<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc [
  <!ENTITY nbsp    "&#160;">
  <!ENTITY zwsp   "&#8203;">
  <!ENTITY nbhy   "&#8209;">
  <!ENTITY wj     "&#8288;">
]>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
<!-- generated by https://github.com/cabo/kramdown-rfc version 1.7.43 (Ruby 3.4.12) -->
<?rfc tocindent="yes"?>
<?rfc strict="yes"?>
<?rfc compact="yes"?>
<?rfc comments="yes"?>
<?rfc inline="yes"?>
<?rfc docmapping="yes"?>
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft-ietf-httpapi-ratelimit-headers-12" category="std" consensus="true" tocInclude="true" sortRefs="true" symRefs="true" version="3">
  <!-- xml2rfc v2v3 conversion 3.35.0 -->
  <front>
    <title>RateLimit header fields for HTTP</title>
    <seriesInfo name="Internet-Draft" value="draft-ietf-httpapi-ratelimit-headers-12"/>
    <author initials="R." surname="Polli" fullname="Roberto Polli">
      <organization>Team Digitale, Italian Government</organization>
      <address>
        <postal>
          <country>Italy</country>
        </postal>
        <email>robipolli@gmail.com</email>
      </address>
    </author>
    <author initials="A." surname="Martinez" fullname="Alejandro Martinez Ruiz">
      <organization>Red Hat</organization>
      <address>
        <email>alex@flawedcode.org</email>
      </address>
    </author>
    <author initials="D." surname="Miller" fullname="Darrel Miller">
      <organization>Microsoft</organization>
      <address>
        <email>darrel@tavis.ca</email>
      </address>
    </author>
    <date year="2026" month="November" day="20"/>
    <area>Applications and Real-Time</area>
    <workgroup>HTTPAPI</workgroup>
    <keyword>Internet-Draft</keyword>
    <abstract>
      <?line 58?>

<t>This document defines the RateLimit-Policy and RateLimit HTTP header fields for servers to advertise their quota policies and the current service limits, thereby allowing clients to shape their requests to avoid exceeding server-imposed limits. It also defines problem types that servers can use to inform clients about quota-related conditions.</t>
    </abstract>
    <note removeInRFC="true">
      <name>About This Document</name>
      <t>
        Status information for this document may be found at <eref target="https://datatracker.ietf.org/doc/draft-ietf-httpapi-ratelimit-headers/"/>.
      </t>
      <t>
        Discussion of this document takes place on the
        HTTPAPI Working Group mailing list (<eref target="mailto:httpapi@ietf.org"/>),
        which is archived at <eref target="https://mailarchive.ietf.org/arch/browse/httpapi/"/>.
        Subscribe at <eref target="https://www.ietf.org/mailman/listinfo/httpapi/"/>.
        Working Group information can be found at <eref target="https://datatracker.ietf.org/wg/httpapi/about/"/>.
      </t>
      <t>Source for this draft and an issue tracker can be found at
        <eref target="https://github.com/ietf-wg-httpapi/ratelimit-headers"/>.</t>
    </note>
  </front>
  <middle>
    <?line 62?>

<section anchor="introduction">
      <name>Introduction</name>
      <t>Rate limiting of HTTP <xref target="HTTP"/> clients is a widespread practice. Servers use quota mechanisms to avoid overload, to ensure an equitable distribution of computational resources, or to enforce other policies such as monetization. Typically, servers limit the number of acceptable requests in a given time window (e.g. 10 requests per second). When a quota is exceeded, servers usually respond with a 4xx HTTP status code (e.g. 429 or 403) or adopt more aggressive policies like dropping connections.</t>
      <t>Quotas may be enforced on different bases (e.g. per user, per IP, per geographic area) and at different levels. To help clients pace their requests, servers can expose quota information via HTTP header fields. This document defines a set of standard HTTP header fields to enable servers to communicate rate limiting information to clients:</t>
      <ul spacing="normal">
        <li>
          <t>RateLimit-Policy: quota policies defined by the server. Client requests consume quota.</t>
        </li>
        <li>
          <t>RateLimit: the quota currently available under specific policies at the time the response is generated.</t>
        </li>
      </ul>
      <t>These fields enable servers to communicate complex rate limiting policies, including using multiple and variable time windows and dynamic quotas, and implementing concurrency limits. This information helps improve the resiliency of HTTP infrastructures by providing clients with the information needed to throttle their requests and avoid 4xx or 5xx responses.</t>
      <t>The following features are out of the scope of this document:</t>
      <dl>
        <dt>Authorization:</dt>
        <dd>
          <t>RateLimit header fields are not meant to support authorization or other kinds of access controls.</t>
        </dd>
        <dt>Response status code:</dt>
        <dd>
          <t>RateLimit header fields may be returned in both successful (see <xref section="15.3" sectionFormat="of" target="HTTP"/>) and non-successful responses. This specification does not cover whether non Successful responses count on quota usage, nor does it mandate any correlation between the RateLimit values and the returned status code.</t>
        </dd>
        <dt>Throttling algorithm:</dt>
        <dd>
          <t>This specification does not mandate a specific throttling algorithm. The values published in the fields, including the window size, can be statically or dynamically evaluated.</t>
        </dd>
        <dt>Service Level Agreement:</dt>
        <dd>
          <t>Conveyed quota hints do not imply any service guarantee. Server is free to throttle clients that adhere to the server’s recommended limits under certain circumstances.</t>
        </dd>
      </dl>
      <section anchor="conventions-and-terminology">
        <name>Conventions and Terminology</name>
        <t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED",
"MAY", and "OPTIONAL" in this document are to be interpreted as
described in BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/> when, and only when, they
appear in all capitals, as shown here.
<?line -6?>
        </t>
        <t>This document uses the terms List, Item, Integer, String, Token, Boolean, Parameters, and Byte Sequence from <xref section="3" sectionFormat="of" target="STRUCTURED-FIELDS"/> to specify syntax and parsing, along with the concept of "bare item".</t>
        <t>The term "problem type" in this document is to be interpreted as described in <xref target="PROBLEM"/>.</t>
        <dl>
          <dt>Quota:</dt>
          <dd>
            <t>A quota is an allocation of capacity used by a server to limit client requests. That capacity is measured in quota units that the client may consume within a time window.</t>
          </dd>
          <dt>Quota Unit:</dt>
          <dd>
            <t>A quota unit is the unit of measurement for the activity of a client.</t>
          </dd>
          <dt>Quota Partition:</dt>
          <dd>
            <t>A quota partition is a division of a server's capacity across different clients, users and owned resources.</t>
          </dd>
          <dt>Time Window:</dt>
          <dd>
            <t>A time window indicates a period of time associated to the allocated quota.</t>
          </dd>
          <dt>Quota Policy:</dt>
          <dd>
            <t>A quota policy is implemented by a server to regulate activity within a specified quota partition, quantified in quota units, over a defined time window. This activity is restricted to a predefined limit, known as the quota. Servers MAY advertise quota policies. More than one quota policy can affect a given request.</t>
          </dd>
          <dt>Service Limit:</dt>
          <dd>
            <t>A service limit is the currently available quota under a specific quota policy and, if defined, the effective time window within which the client can use no more than the available quota.</t>
          </dd>
        </dl>
      </section>
    </section>
    <section anchor="ratelimit-policy-field">
      <name>RateLimit-Policy Field</name>
      <t>The "RateLimit-Policy" response header field is a List whose members are Tokens or Strings. Each list item identifies a Quota Policy (<xref target="quotapolicy-item"/>).</t>
      <t>This field MUST NOT appear in a trailer section.</t>
      <t>A recipient MUST ignore a RateLimit-Policy field received in a trailer section.</t>
      <t>The "RateLimit-Policy" field enables clients to pace their requests based on policy information provided by the server. Servers SHOULD keep the field value consistent over a sequence of HTTP responses because changes can prevent clients from correctly following the communicated policy. Situations where throttling constraints are highly dynamic are better served using the RateLimit field (<xref target="ratelimit-field"/>) that communicates the latest service information a client can react to. Both fields can be communicated by the server when appropriate.</t>
      <artwork><![CDATA[
   RateLimit-Policy: burst;q=100;w=60,"daily";q=1000;w=86400
]]></artwork>
      <section anchor="quotapolicy-item">
        <name>Quota Policy Item</name>
        <t>A Quota Policy is identified by a Token or String. A set of Parameters describe mandatory and optional details of the policy.</t>
        <t>Policy identifiers are compared case-insensitively by their character value. The bare item type is not part of the identifier. For example, <tt>daily</tt>, <tt>DAILY</tt>, and <tt>"daily"</tt> all identify the same policy. Servers MUST NOT generate multiple items in the same field whose policy identifiers are equivalent under this comparison.</t>
        <t>The following parameters are defined:</t>
        <dl>
          <dt>q:</dt>
          <dd>
            <t>The REQUIRED "q" parameter indicates the quota allocated by this policy measured in quota units (<xref target="ratelimitpolicy-quota"/>).</t>
          </dd>
          <dt>qu:</dt>
          <dd>
            <t>The OPTIONAL "qu" parameter value conveys the quota units associated to the "q" parameter (<xref target="ratelimitpolicy-quotaunit"/>). The default quota unit is "requests".</t>
          </dd>
          <dt>w:</dt>
          <dd>
            <t>The OPTIONAL "w" parameter value conveys a time window (<xref target="ratelimitpolicy-window"/>).</t>
          </dd>
          <dt>pk:</dt>
          <dd>
            <t>The OPTIONAL "pk" parameter value conveys the partition key associated to the corresponding request (<xref target="ratelimitpolicy-partitionkey"/>).</t>
          </dd>
        </dl>
        <t>Unknown parameters MUST be ignored by recipients. A recipient that receives a RateLimit-Policy field with a parameter value that does not conform to the syntax defined in this section SHOULD ignore the malformed list item.</t>
        <section anchor="ratelimitpolicy-quota">
          <name>Quota Parameter</name>
          <t>The "q" parameter value MUST be a non-negative Integer. The value indicates the quota allocated for client activity (measured in quota units) for a given quota partition.</t>
        </section>
        <section anchor="ratelimitpolicy-quotaunit">
          <name>Quota Unit Parameter</name>
          <t>The "qu" parameter value conveys the quota units applicable to the quota (<xref target="ratelimitpolicy-quota"/>). The value MUST be a String. Allowed values are listed in the <xref target="ratelimit-quota-unit-registry">RateLimit Quota Units registry</xref>. This specification defines three quota units:</t>
          <dl>
            <dt>requests:</dt>
            <dd>
              <t>This value indicates the quota is based on the number of requests processed by the server. Whether a specific request actually consumes a quota unit is implementation-specific.</t>
            </dd>
            <dt>content-bytes:</dt>
            <dd>
              <t>This value indicates the quota is based on the number of content bytes processed by the server.</t>
            </dd>
            <dt>concurrent-requests:</dt>
            <dd>
              <t>This value indicates the quota is based on the number of concurrent requests processed by the server.</t>
            </dd>
          </dl>
        </section>
        <section anchor="ratelimitpolicy-window">
          <name>Window Parameter</name>
          <t>The "w" parameter value conveys a time window applicable to the quota (<xref target="ratelimitpolicy-quota"/>). The time window MUST be a non-negative, non-zero, Integer value expressing an interval in seconds, similar to the "delay-seconds" rule defined in <xref section="10.2.3" sectionFormat="of" target="HTTP"/>. Sub-second precision is not supported.</t>
        </section>
        <section anchor="ratelimitpolicy-partitionkey">
          <name>Partition Key Parameter</name>
          <t>The "pk" parameter value conveys the partition key associated to the request. The value MUST be a Byte Sequence. Servers MAY use the partition key to divide server capacity across different clients and resources. Quotas are allocated per partition key.</t>
        </section>
      </section>
    </section>
    <section anchor="ratelimit-field">
      <name>RateLimit Field</name>
      <t>The "RateLimit" response header field is a List whose members are Tokens or Strings. Each list item identifies the current service limit for a quota policy.</t>
      <t>This field MUST NOT appear in a trailer section.</t>
      <t>A recipient MUST ignore a RateLimit field received in a trailer section.</t>
      <t>A server uses this field to communicate the remaining quota for a specific policy at the time the response is generated.</t>
      <artwork><![CDATA[
   RateLimit: default;a=50;w=30
]]></artwork>
      <section anchor="servicelimit-item">
        <name>Service Limit Item</name>
        <t>Each Service Limit Item identifies the quota policy (<xref target="quotapolicy-item"/>) associated with the request and contains parameters with information about the current service limit.</t>
        <t>The following parameters are defined in this specification:</t>
        <dl>
          <dt>a:</dt>
          <dd>
            <t>This REQUIRED parameter value conveys the available quota under the identified policy (<xref target="ratelimit-available-quota"/>).</t>
          </dd>
          <dt>w:</dt>
          <dd>
            <t>This OPTIONAL parameter value conveys the effective window under the identified policy: the time within which the client can use no more than the available quota (<xref target="ratelimit-effective-window"/>).</t>
          </dd>
          <dt>pk:</dt>
          <dd>
            <t>The OPTIONAL "pk" parameter value conveys the partition key associated to the corresponding request (<xref target="ratelimit-partitionkey"/>).</t>
          </dd>
          <dt>c:</dt>
          <dd>
            <t>The OPTIONAL "c" parameter value conveys the cost of the corresponding request in quota units (<xref target="ratelimit-cost"/>).</t>
          </dd>
        </dl>
        <t>Unknown parameters MUST be ignored by recipients.</t>
        <section anchor="ratelimit-available-quota">
          <name>Available Quota Parameter</name>
          <t>The "a" parameter indicates the available quota under the identified policy.</t>
          <t>It is a non-negative Integer expressed in quota units.
Clients MUST NOT assume that a positive available quota is a guarantee that further requests will be served.
When the available quota is low, it indicates that the server may soon throttle the client (see <xref target="providing-ratelimit-fields"/>).</t>
        </section>
        <section anchor="ratelimit-effective-window">
          <name>Effective Window Parameter</name>
          <t>The "w" parameter indicates effective window under the identified policy: the number of seconds within which the client can use no more than the available quota.</t>
          <t>It is a non-negative Integer compatible with the delay-seconds rule, because:</t>
          <ul spacing="normal">
            <li>
              <t>it does not rely on clock synchronization and is resilient to clock adjustment
and clock skew between client and server (see <xref section="5.6.7" sectionFormat="of" target="HTTP"/>);</t>
            </li>
            <li>
              <t>it mitigates the risk related to thundering herd when too many clients are serviced with the same timestamp.</t>
            </li>
          </ul>
          <t>The client MUST NOT assume that all its service limit will be fully restored after the time indicated by the effective window parameter. The server MAY arbitrarily alter the available quota and effective window between subsequent requests; for example, in case of resource saturation or to implement sliding window policies.</t>
        </section>
        <section anchor="ratelimit-partitionkey">
          <name>Partition Key Parameter</name>
          <t>The "pk" parameter value conveys the partition key associated to the request. The value MUST be a Byte Sequence. Servers MAY use the partition key to divide server capacity across different clients and resources. Quotas are allocated per partition key.</t>
        </section>
        <section anchor="ratelimit-cost">
          <name>Cost Parameter</name>
          <t>The "c" parameter indicates the cost of the corresponding request in quota units. It is a non-negative Integer. When present, it enables a client to estimate how many similar requests it can make within the available quota without relying on the difference between successive available quota values, which may be unreliable when requests are made concurrently or when the server adjusts quota between responses.</t>
        </section>
      </section>
    </section>
    <section anchor="ratelimit-partition-field">
      <name>RateLimit-Partition Field</name>
      <t>The "RateLimit-Partition" response header field is a List whose members are Tokens or Strings. Each list item identifies a quota policy and declares the dimensions used to partition quota for that policy.</t>
      <t>This field MUST NOT appear in a trailer section.</t>
      <t>A recipient MUST ignore a RateLimit-Partition field received in a trailer section.</t>
      <artwork><![CDATA[
   RateLimit-Partition: api;user_id;method
]]></artwork>
      <section anchor="partition-dimension-item">
        <name>Partition Dimension Item</name>
        <t>Each item in the RateLimit-Partition field is a Token or String that matches, as described in <xref target="quotapolicy-item"/>, the policy identifier used in the corresponding RateLimit-Policy (<xref target="ratelimit-policy-field"/>) and RateLimit (<xref target="ratelimit-field"/>) fields. The item's Parameters declare the partition dimensions.</t>
        <t>Each parameter key names a partition dimension registered in the RateLimit Partition Key Dimensions registry (<xref target="ratelimit-partition-key-dimensions-registry"/>).</t>
        <t>A dimension parameter with a Boolean value of true (e.g. <tt>;user_id</tt>) indicates that the dimension varies across partitions — the server uses the request's value for that dimension when constructing the partition key.</t>
        <t>A dimension parameter with a non-Boolean value (e.g. <tt>;method=GET</tt>) indicates that the policy applies only to requests where that dimension matches the specified value. This enables a client to determine whether a policy is relevant to a given request before it is sent.</t>
        <t>Unknown dimension parameters SHOULD be ignored by recipients when processing the field, but a client that does not recognize or cannot compute the value of any dimension MUST treat the partition key for that policy as opaque and MUST NOT attempt to construct partition keys for prediction purposes.</t>
      </section>
      <section anchor="partition-key-construction">
        <name>Partition Key Construction</name>
        <t>When a server provides both the RateLimit-Partition field and a "pk" parameter in the RateLimit field, the "pk" Byte Sequence MUST be constructed as follows:</t>
        <ol spacing="normal" type="1"><li>
            <t>Collect the dimension parameters from the RateLimit-Partition item. For dimensions with a Boolean value of true, the value is the request's value for that dimension. For dimensions with a non-Boolean value, the value is the parameter value itself.</t>
          </li>
          <li>
            <t>Sort the dimensions by parameter key in lexicographic order.</t>
          </li>
          <li>
            <t>For each dimension in sorted order, encode the value as a UTF-8 byte string.</t>
          </li>
          <li>
            <t>Concatenate the encoded values in order, separated by the byte 0x1F (Unit Separator, US).</t>
          </li>
          <li>
            <t>The resulting byte string is the partition key Byte Sequence.</t>
          </li>
        </ol>
        <t>Dimension values MUST NOT contain the byte 0x1F. If a dimension value would naturally contain this byte, the server MUST define an escaping or encoding mechanism in the dimension's registry entry.</t>
        <t>For example, if the server declares:</t>
        <artwork><![CDATA[
   RateLimit-Partition: "api";user_id;method
]]></artwork>
        <t>and the request is made by user "alice" using the GET method, the partition key is constructed by sorting the dimensions alphabetically (method, user_id), then concatenating "GET", 0x1F, and "alice". The RateLimit field conveys this as a Byte Sequence:</t>
        <artwork><![CDATA[
   RateLimit: "api";a=99;w=60;pk=:R0VUH2FsaWNl:
]]></artwork>
        <t>A server might also declare a policy that applies only to a specific method:</t>
        <artwork><![CDATA[
   RateLimit-Partition: "reads";user_id;method=GET
]]></artwork>
        <t>A client receiving this knows the "reads" policy only applies to GET requests. For non-GET requests, this policy is not relevant and can be ignored.</t>
        <t>A client that knows its own values for the declared dimensions can construct the expected partition key for a future request and look it up in its local quota tracking state, enabling proactive traffic shaping without requiring API-specific knowledge.</t>
        <t>If a server provides a "pk" parameter without a corresponding RateLimit-Partition field item, the structure of the partition key is not defined by this specification and any interpretation of its structure MUST be communicated out of band. Clients MAY use heuristics to estimate partition membership based on observed partition key values and request similarity.</t>
      </section>
    </section>
    <section anchor="providing-ratelimit-fields">
      <name>Server Behavior</name>
      <t>A server MAY return RateLimit header fields independently of the response status code. This includes throttled responses. This document does not mandate any correlation between the RateLimit header field values and the returned status code.</t>
      <t>Servers SHOULD avoid returning RateLimit header fields in redirection responses (i.e., responses with 3xx status codes) because a low available quota could prevent the client from following the redirect. For example, given the RateLimit header fields below, a client could decide to wait 10 seconds before following the "Location" header field (see <xref section="10.2.2" sectionFormat="of" target="HTTP"/>), because the available quota is 0.</t>
      <sourcecode type="http-message"><![CDATA[
HTTP/1.1 301 Moved Permanently
Location: /foo/123
RateLimit: "problemPolicy";a=0;w=10

]]></sourcecode>
      <t>If a response contains both the Retry-After and the RateLimit header fields, the Retry-After field value SHOULD NOT reference a point in time earlier than the end of the effective window.</t>
      <t>Servers using RateLimit header fields SHOULD implement mechanisms to cap the ratio between the available quota and the effective window values (see <xref target="sec-resource-exhaustion"/>); this is especially important when a quota policy uses a large time window, as a large available quota relative to a small effective window could inadvertently encourage a burst of requests.</t>
      <t>Under certain conditions, a server MAY reduce the available quota or increase the effective window between subsequent requests, e.g. to respond to Denial of Service attacks or in case of resource saturation.</t>
      <section anchor="performance-considerations">
        <name>Performance Considerations</name>
        <t>Servers are not required to return RateLimit header fields in every response, and clients need to take this into account. For example, an implementer concerned with performance might provide RateLimit header fields only when a given quota is close to exhaustion.</t>
        <t>Implementers concerned with response fields' size, might take into account their ratio with respect to the content length, or use header-compression HTTP features such as <xref target="HPACK"/>.</t>
      </section>
    </section>
    <section anchor="quota-partitions">
      <name>Quota Partitions</name>
      <t>Servers commonly enforce different quota allocations depending on attributes of the request — for example, per authenticated user, per client application, per HTTP method, or per resource. When a single client makes requests that span multiple such allocations, the client needs to track quota for each allocation independently in order to shape its traffic effectively.</t>
      <t>A key goal of partition keys is to enable generic client-side traffic shaping without requiring API-specific knowledge. Historically, client-side rate limiting code has relied on out-of-band documentation to understand how a server partitions its quota — for example, by reading API documentation that explains rate limits are per-user or per-endpoint. This prevents the development of generic HTTP client libraries that can automatically shape traffic for any API.</t>
      <t>Partition keys address this by allowing the server to communicate the dimensions along which quota is partitioned, using attributes that the client can observe from its own requests. When a client can determine which partition a future request will fall into, it can proactively check whether sufficient quota is available before sending the request, rather than waiting to be throttled.</t>
      <t>The RateLimit-Partition field (<xref target="ratelimit-partition-field"/>) enables servers to declare the named dimensions that make up the partition key for each policy. When this field is provided, clients can construct partition keys deterministically from their own request attributes and predict which quota partition a future request will consume.</t>
      <t>For example, a server that allocates separate quotas per user and HTTP method might declare:</t>
      <artwork><![CDATA[
   RateLimit-Partition: "api";user_id;method
]]></artwork>
      <t>Partition keys returned by the server in RateLimit header fields allow a single client to track the quota of multiple partitions independently. This is useful when a single client makes requests that consume different quota allocations, for example when acting on behalf of different users or accessing different resources that have independent quotas.</t>
      <t>Servers SHOULD construct partition keys using only information that is present in the request. Servers SHOULD avoid returning partition keys that contain sensitive information. Where partition keys contain identifying information, either of the client application or the user, servers should be aware of the potential for impersonation and apply the appropriate security mechanisms.</t>
      <t>For cases where the server does not provide a RateLimit-Partition field, or where a client does not recognize all declared dimensions, clients MAY use heuristics to estimate whether a future request will be successful based on its similarity to previous requests.</t>
    </section>
    <section anchor="receiving-fields">
      <name>Client Behavior</name>
      <t>The RateLimit header fields can be used by clients to determine whether the associated request respected the server's quota policy, and as an indication of whether subsequent requests will be successful. However, the server might apply other criteria when servicing future requests, and so the quota policy may not completely reflect whether requests will succeed.</t>
      <t>For example, a successful response with the following fields:</t>
      <artwork><![CDATA[
   RateLimit: "default";a=1;w=7
]]></artwork>
      <t>does not guarantee that the next request will be successful. Servers' behavior may be subject to other conditions.</t>
      <t>A client is responsible for ensuring that RateLimit header field values returned
cause reasonable client behavior with respect to throughput and latency
(see <xref target="sec-resource-exhaustion"/> and <xref target="sec-dos"/>).</t>
      <t>A client receiving RateLimit header fields MUST NOT assume that future responses will contain the same RateLimit header fields, or any RateLimit header fields at all.</t>
      <t>Malformed RateLimit header fields MUST be ignored.</t>
      <t>A client SHOULD NOT exceed the available quota within the time expressed in the effective window parameter.</t>
      <t>The value of the effective window parameter is generated at response time: a client aware of a significant network latency MAY behave accordingly and use other information (e.g. the "Date" response header field, or otherwise gathered metrics) to schedule requests.</t>
      <t>The details provided in the RateLimit-Policy header field are informative and MAY be ignored.</t>
      <t>If a response contains both the RateLimit and Retry-After fields, the Retry-After field MUST take precedence and the effective window MAY be ignored.</t>
      <t>This specification does not mandate a specific throttling behavior and implementers can adopt their preferred policies, including:</t>
      <ul spacing="normal">
        <li>
          <t>slowing down or pre-emptively back-off their request rate when
approaching quota limits;</t>
        </li>
        <li>
          <t>consuming all the quota according to the exposed limits and then wait.</t>
        </li>
      </ul>
      <section anchor="intermediaries">
        <name>Intermediaries</name>
        <t>This section documents the considerations advised in <xref section="16.3.2" sectionFormat="of" target="HTTP"/>.</t>
        <t>An intermediary that is not part of the originating service infrastructure and is not aware of the quota policy semantics used by the origin server SHOULD NOT alter the RateLimit header fields' values in such a way as to communicate a more permissive quota policy; this includes removing the RateLimit header fields.</t>
        <t>An intermediary that is aware of the quota unit semantics used by the origin server MAY alter the RateLimit header fields to communicate a different quota policy, for example when the intermediary enforces its own quota policy or when it can absorb additional load through caching or request collapsing.</t>
        <t>An intermediary SHOULD forward a request even when presuming that it might not be serviced; the service returning the RateLimit header fields is solely responsible for enforcing the communicated quota policy, and MAY still choose to service incoming requests.</t>
        <t>This specification does not mandate any behavior on intermediaries with respect to retries, nor does it require that intermediaries have any role in respecting quota policies. For example, it is legitimate for a proxy to retransmit a request without notifying the client, and thus consuming quota units.</t>
        <t><xref target="privacy">Privacy considerations</xref> provide further guidance on intermediaries.</t>
      </section>
    </section>
    <section anchor="problem-types">
      <name>Problem Types</name>
      <section anchor="quota-exceeded">
        <name>Quota Exceeded</name>
        <t>This section defines the "https://iana.org/assignments/http-problem-types#quota-exceeded" problem type. A server MAY use this problem type if it wants to communicate to the client that the requests sent by the client exceed one or more Quota Policies. This problem type defines the extension member "violated-policies" as an array of strings, whose value is the names of policies where the quota was exceeded.</t>
        <sourcecode type="http-message"><![CDATA[
HTTP/1.1 429 Too Many Requests
Content-Type: application/problem+json

{
  "type": "https://iana.org/assignments/http-problem-types#quota-exceeded",
  "title": "Request cannot be satisfied as assigned quota has been exceeded",
  "violated-policies": ["daily","bandwidth"]
}
]]></sourcecode>
      </section>
      <section anchor="temporary-reduced-capacity">
        <name>Temporary Reduced Capacity</name>
        <t>This section defines the "https://iana.org/assignments/http-problem-types#temporary-reduced-capacity" problem type. A server MAY use this problem type if it wants to communicate that the client’s requests currently cannot be satisfied due to a temporary reduction in server capacity. The server MAY choose to include a RateLimit-Policy field indicating the new temporarily lower quota. This problem type defines the extension member "violated-policies" as an array of strings, whose value is the names of policies where the quota was exceeded.</t>
        <sourcecode type="http-message"><![CDATA[
HTTP/1.1 503 Service Unavailable
Content-Type: application/problem+json

{
  "type": "https://iana.org/assignments/http-problem-types#temporary-reduced-capacity",
  "title": "Request cannot be satisfied due to temporary server capacity constraints",
  "violated-policies": ["hourly"]
}
]]></sourcecode>
      </section>
      <section anchor="abnormal-usage-detected">
        <name>Abnormal Usage Detected</name>
        <t>This section defines the "https://iana.org/assignments/http-problem-types#abnormal-usage-detected" problem type. A server MAY use this problem type  to communicate to the client that it has detected a pattern of requests that suggest unintentional or malicious behaviour on the part of the client. This problem type defines the extension member "violated-policies" as an array of strings, whose value is the names of policies where the quota was exceeded.</t>
        <sourcecode type="http-message"><![CDATA[
HTTP/1.1 429 Too Many Requests
Content-Type: application/problem+json

{
  "type": "https://iana.org/assignments/http-problem-types#abnormal-usage-detected",
  "title": "Request not satisifed due to detection of abnormal request pattern",
  "violated-policies": ["hourly"]
}
]]></sourcecode>
      </section>
      <section anchor="caching">
        <name>Caching</name>
        <t><xref target="HTTP-CACHING"/> defines how responses can be stored and reused for subsequent requests,
including those with RateLimit header fields.
Because the information in RateLimit header fields on a cached response may not be current, they SHOULD be ignored on responses that come from cache
(i.e., those with a positive current_age; see <xref section="4.2.3" sectionFormat="of" target="HTTP-CACHING"/>).</t>
      </section>
    </section>
    <section anchor="security-considerations">
      <name>Security Considerations</name>
      <section anchor="sec-throttling-does-not-prevent">
        <name>Throttling does not prevent clients from issuing requests</name>
        <t>This specification does not prevent clients from making requests.
Servers should always implement mechanisms to prevent resource exhaustion.</t>
      </section>
      <section anchor="sec-information-disclosure">
        <name>Information disclosure</name>
        <t>Servers should not disclose to untrusted parties operational capacity information
that can be used to saturate its infrastructural resources.</t>
        <t>While this specification does not mandate whether non-successful responses consume quota,
if error responses (such as 401 (Unauthorized) and 403 (Forbidden)) count against quota,
a malicious client could probe the endpoint to get traffic information of another user.</t>
        <t>As intermediaries might retransmit requests and consume
quota units without prior knowledge of the user agent,
RateLimit header fields might reveal the existence of an intermediary
to the user agent.</t>
        <t>Partition keys may contain identifying information; see <xref target="quota-partitions"/> for guidance on constructing partition keys.</t>
      </section>
      <section anchor="sec-available-not-granted">
        <name>Available quota units are not granted requests</name>
        <t>RateLimit header fields convey hints from the server
to the clients in order to help them avoid being throttled out.</t>
        <t>Clients MUST NOT consider the <xref target="ratelimit-available-quota">available quota parameter</xref> as a service level agreement.</t>
        <t>In case of resource saturation, the server MAY artificially lower the returned values
or not serve the request regardless of the advertised quotas.</t>
      </section>
      <section anchor="sec-window-variability">
        <name>Variability of the effective window</name>
        <t>The <xref target="ratelimit-effective-window">effective window parameter</xref> does not imply anything about when the server will increase the available quota. The effective window does not necessarily end at a fixed point in time.</t>
        <t>Subsequent requests might return a higher effective window value
to limit concurrency or implement dynamic or adaptive throttling policies.</t>
      </section>
      <section anchor="sec-resource-exhaustion">
        <name>Resource exhaustion</name>
        <t>When returning effective window values, servers must be aware that
many throttled clients may come back at the very moment specified.</t>
        <t>This is true for Retry-After too.</t>
        <t>For example, if the quota resets every day at <tt>18:00:00</tt>
and your server returns the effective window accordingly</t>
        <artwork><![CDATA[
   Date: Tue, 15 Nov 1994 18:00:00 GMT
   RateLimit: "daily";a=1;w=86400
]]></artwork>
        <t>there's a high probability that all clients will show up at <tt>18:00:00</tt>.</t>
        <t>This could be mitigated by adding some jitter to the effective window.</t>
        <t>Resource exhaustion issues can be associated with quota policies using a
large time window, because a user agent by chance or on purpose
might consume most of its quota units in a significantly shorter interval.</t>
        <t>This behavior can be even triggered by the provided RateLimit header fields.
The following example describes a service
with an unconsumed quota policy of 10000 quota units per 1000 seconds.</t>
        <artwork><![CDATA[
RateLimit-Policy: "somepolicy";q=10000;w=1000
RateLimit: "somepolicy";a=10000;w=10
]]></artwork>
        <t>A client implementing a simple ratio between available quota and effective window could infer an average throughput of 1000 quota units per second, while the policy's quota and window parameters convey an average of 10 quota units per second.
If the service cannot handle such load, it should return either a lower available quota or a higher effective window.
Moreover, complementing large time window quota policies with a short time window one mitigates those risks.</t>
        <section anchor="sec-dos">
          <name>Denial of Service</name>
          <t>RateLimit header fields may contain unexpected values by chance or on purpose.
For example, an excessively high available quota may be:</t>
          <ul spacing="normal">
            <li>
              <t>used by a malicious intermediary to trigger a Denial of Service attack
or consume client resources boosting its requests;</t>
            </li>
            <li>
              <t>passed by a misconfigured server;</t>
            </li>
          </ul>
          <t>or a large effective window could inhibit clients to contact the server (e.g. similarly to receiving "Retry-after: 1000000").</t>
          <t>To mitigate this risk, clients can set thresholds that they consider reasonable in terms of quota units, time window, concurrent requests or throughput, and define a consistent behavior when the RateLimit exceed those thresholds.
For example this means capping the maximum number of request per second, or implementing retries when the effective window exceeds ten minutes.</t>
          <t>The considerations above are not limited to RateLimit header fields, but apply to all fields affecting how clients behave in subsequent requests (e.g. Retry-After).</t>
        </section>
      </section>
    </section>
    <section anchor="privacy">
      <name>Privacy Considerations</name>
      <t>Clients that act upon a request to rate limit
are potentially re-identifiable (see <xref section="5.2.1" sectionFormat="of" target="PRIVACY"/>)
because they react to information that might only be given to them.
Note that this might apply to other fields too (e.g. Retry-After).</t>
      <t>Since rate limiting is usually implemented in contexts where
clients are either identified or profiled
(e.g. assigning different quota units to different users),
this is rarely a concern.</t>
      <t>Privacy enhancing infrastructures using RateLimit header fields
can define specific techniques to mitigate the risks of re-identification.</t>
    </section>
    <section anchor="iana-considerations">
      <name>IANA Considerations</name>
      <t>IANA is requested to update two registries and create five new registries.</t>
      <section anchor="update-http-field-name-registry">
        <name>Update HTTP Field Name Registry</name>
        <t>Please add the following entries to the
"Hypertext Transfer Protocol (HTTP) Field Name Registry" registry (<xref target="HTTP"/>):</t>
        <table>
          <thead>
            <tr>
              <th align="left">Field Name</th>
              <th align="left">Structured Type</th>
              <th align="left">Status</th>
              <th align="left">Specification</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td align="left">RateLimit</td>
              <td align="left">List</td>
              <td align="left">permanent</td>
              <td align="left">
                <xref target="ratelimit-field"/> of RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">RateLimit-Policy</td>
              <td align="left">List</td>
              <td align="left">permanent</td>
              <td align="left">
                <xref target="ratelimit-policy-field"/> of RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">RateLimit-Partition</td>
              <td align="left">List</td>
              <td align="left">permanent</td>
              <td align="left">
                <xref target="ratelimit-partition-field"/> of RFC nnnn</td>
            </tr>
          </tbody>
        </table>
      </section>
      <section anchor="update-http-problem-type-registry">
        <name>Update HTTP Problem Type registry</name>
        <t>IANA is asked to register the following entries in the "HTTP Problem Types" registry at https://www.iana.org/assignments/http-problem-types.</t>
        <section anchor="registration-of-quota-exceeded-problem-type">
          <name>Registration of "quota-exceeded" Problem Type</name>
          <t>Type URI: https://iana.org/assignments/http-problem-types#quota-exceeded</t>
          <t>Title: Quota Exceeded</t>
          <t>Recommended HTTP status code: 429</t>
          <t>Reference: <xref target="quota-exceeded"/> of this document</t>
        </section>
        <section anchor="registration-of-temporary-reduced-capacity-problem-type">
          <name>Registration of "temporary-reduced-capacity" Problem Type</name>
          <t>Type URI: https://iana.org/assignments/http-problem-types#temporary-reduced-capacity</t>
          <t>Title: Temporary Reduced Capacity</t>
          <t>Recommended HTTP status code: 503</t>
          <t>Reference: <xref target="temporary-reduced-capacity"/> of this document</t>
        </section>
        <section anchor="registration-of-abnormal-usage-detected-problem-type">
          <name>Registration of "abnormal-usage-detected" Problem Type</name>
          <t>Type URI: https://iana.org/assignments/http-problem-types#abnormal-usage-detected</t>
          <t>Title: Abnormal Usage Detected</t>
          <t>Recommended HTTP status code: 429</t>
          <t>Reference: <xref target="abnormal-usage-detected"/> of this document</t>
        </section>
      </section>
      <section anchor="ratelimit-quota-unit-registry">
        <name>RateLimit quota unit registry</name>
        <t>This specification establishes the registry "Hypertext Transfer Protocol (HTTP) RateLimit Quota Units" registry to be located at https://www.iana.org/assignments/http-ratelimit-quota-units. Registration is done on the advice of a Designated Expert, appointed by the IESG or their delegate. All entries are Specification Required (<xref target="IANA"/>, Section 4.6).</t>
        <t>The registry has the following initial content:</t>
        <table>
          <thead>
            <tr>
              <th align="left">Quota Unit</th>
              <th align="left">Reference</th>
              <th align="left">Notes</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td align="left">request</td>
              <td align="left">RFC nnnn</td>
              <td align="left"> </td>
            </tr>
            <tr>
              <td align="left">content-bytes</td>
              <td align="left">RFC nnnn</td>
              <td align="left"> </td>
            </tr>
            <tr>
              <td align="left">concurrent-requests</td>
              <td align="left">RFC nnnn</td>
              <td align="left"> </td>
            </tr>
          </tbody>
        </table>
        <section anchor="registration-template">
          <name>Registration Template</name>
          <t>The registration template for the RateLimit Quota Units registry is as follows:</t>
          <ul spacing="normal">
            <li>
              <t>Quota Unit: The name of the quota unit.</t>
            </li>
            <li>
              <t>Reference: A reference to the document that specifies the quota unit.</t>
            </li>
            <li>
              <t>Notes: Any additional notes about the quota unit.</t>
            </li>
          </ul>
        </section>
      </section>
      <section anchor="ratelimit-policy-parameters-registry">
        <name>RateLimit-Policy parameters registry</name>
        <t>This specification establishes the registry "Hypertext Transfer Protocol (HTTP) RateLimit-Policy Parameters" to be located at https://www.iana.org/assignments/http-ratelimit-policy-parameters. Registration is done on the advice of a Designated Expert, appointed by the IESG or their delegate. All entries are Specification Required (<xref target="IANA"/>, Section 4.6).</t>
        <t>Each registered parameter defines a named parameter that may appear on items in the RateLimit-Policy header field. Parameters not registered in this registry SHOULD use a vendor prefix (e.g. <tt>acme-burst</tt>) to avoid collisions.</t>
        <t>The registry has the following initial content:</t>
        <table>
          <thead>
            <tr>
              <th align="left">Parameter</th>
              <th align="left">Description</th>
              <th align="left">Reference</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td align="left">q</td>
              <td align="left">Quota allocated by the policy</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">qu</td>
              <td align="left">Quota units for the policy</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">w</td>
              <td align="left">Time window for the policy</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">pk</td>
              <td align="left">Partition key</td>
              <td align="left">RFC nnnn</td>
            </tr>
          </tbody>
        </table>
        <section anchor="registration-template-1">
          <name>Registration Template</name>
          <t>The registration template for the RateLimit-Policy Parameters registry is as follows:</t>
          <ul spacing="normal">
            <li>
              <t>Parameter: The parameter key. The key MUST conform to the syntax for Structured Fields parameter keys.</t>
            </li>
            <li>
              <t>Description: A description of the parameter's semantics.</t>
            </li>
            <li>
              <t>Reference: A reference to the document that specifies the parameter.</t>
            </li>
          </ul>
        </section>
      </section>
      <section anchor="ratelimit-parameters-registry">
        <name>RateLimit parameters registry</name>
        <t>This specification establishes the registry "Hypertext Transfer Protocol (HTTP) RateLimit Parameters" to be located at https://www.iana.org/assignments/http-ratelimit-parameters. Registration is done on the advice of a Designated Expert, appointed by the IESG or their delegate. All entries are Specification Required (<xref target="IANA"/>, Section 4.6).</t>
        <t>Each registered parameter defines a named parameter that may appear on items in the RateLimit header field. Parameters not registered in this registry SHOULD use a vendor prefix (e.g. <tt>acme-burst</tt>) to avoid collisions.</t>
        <t>The registry has the following initial content:</t>
        <table>
          <thead>
            <tr>
              <th align="left">Parameter</th>
              <th align="left">Description</th>
              <th align="left">Reference</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td align="left">a</td>
              <td align="left">Available quota</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">w</td>
              <td align="left">Effective window</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">pk</td>
              <td align="left">Partition key</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">c</td>
              <td align="left">Cost of the request in quota units</td>
              <td align="left">RFC nnnn</td>
            </tr>
          </tbody>
        </table>
        <section anchor="registration-template-2">
          <name>Registration Template</name>
          <t>The registration template for the RateLimit Parameters registry is as follows:</t>
          <ul spacing="normal">
            <li>
              <t>Parameter: The parameter key. The key MUST conform to the syntax for Structured Fields parameter keys.</t>
            </li>
            <li>
              <t>Description: A description of the parameter's semantics.</t>
            </li>
            <li>
              <t>Reference: A reference to the document that specifies the parameter.</t>
            </li>
          </ul>
        </section>
      </section>
      <section anchor="ratelimit-partition-parameters-registry">
        <name>RateLimit-Partition parameters registry</name>
        <t>This specification establishes the registry "Hypertext Transfer Protocol (HTTP) RateLimit-Partition Parameters" to be located at https://www.iana.org/assignments/http-ratelimit-partition-parameters. Registration is done on the advice of a Designated Expert, appointed by the IESG or their delegate. All entries are Specification Required (<xref target="IANA"/>, Section 4.6).</t>
        <t>Each registered parameter defines a named parameter that may appear on items in the RateLimit-Partition header field. Partition dimension parameters are registered separately in the RateLimit Partition Key Dimensions registry (<xref target="ratelimit-partition-key-dimensions-registry"/>). Other parameters not registered in this registry SHOULD use a vendor prefix to avoid collisions.</t>
        <t>The registry has no initial content beyond the partition key dimensions defined in <xref target="ratelimit-partition-key-dimensions-registry"/>.</t>
        <section anchor="registration-template-3">
          <name>Registration Template</name>
          <t>The registration template for the RateLimit-Partition Parameters registry is as follows:</t>
          <ul spacing="normal">
            <li>
              <t>Parameter: The parameter key. The key MUST conform to the syntax for Structured Fields parameter keys.</t>
            </li>
            <li>
              <t>Description: A description of the parameter's semantics.</t>
            </li>
            <li>
              <t>Reference: A reference to the document that specifies the parameter.</t>
            </li>
          </ul>
        </section>
      </section>
      <section anchor="ratelimit-partition-key-dimensions-registry">
        <name>RateLimit partition key dimensions registry</name>
        <t>This specification establishes the registry "Hypertext Transfer Protocol (HTTP) RateLimit Partition Key Dimensions" registry to be located at https://www.iana.org/assignments/http-ratelimit-partition-key-dimensions. Registration is done on the advice of a Designated Expert, appointed by the IESG or their delegate. All entries are Specification Required (<xref target="IANA"/>, Section 4.6).</t>
        <t>Each registered dimension defines a named attribute of an HTTP request that a server may use to partition quota allocations. For a dimension to be useful for client-side traffic shaping, the client MUST be able to determine the dimension value from information available before sending the request.</t>
        <t>The registry has the following initial content:</t>
        <table>
          <thead>
            <tr>
              <th align="left">Dimension</th>
              <th align="left">Description</th>
              <th align="left">Reference</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td align="left">user_id</td>
              <td align="left">The identifier of the user on whose behalf the request is made, as known to the server.</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">client_id</td>
              <td align="left">The identifier of the client application making the request, as known to the server.</td>
              <td align="left">RFC nnnn</td>
            </tr>
            <tr>
              <td align="left">method</td>
              <td align="left">The HTTP method of the request, encoded as its canonical uppercase form (e.g. "GET", "POST").</td>
              <td align="left">RFC nnnn</td>
            </tr>
          </tbody>
        </table>
        <section anchor="registration-template-4">
          <name>Registration Template</name>
          <t>The registration template for the RateLimit Partition Key Dimensions registry is as follows:</t>
          <ul spacing="normal">
            <li>
              <t>Dimension: The name of the partition key dimension. The name MUST conform to the syntax for Structured Fields Tokens, starting with a lowercase letter.</t>
            </li>
            <li>
              <t>Description: A description of the dimension, including the request attribute it corresponds to and any canonicalization or normalization rules for the value.</t>
            </li>
            <li>
              <t>Reference: A reference to the document that specifies the dimension.</t>
            </li>
            <li>
              <t>Notes: Any additional notes about the dimension.</t>
            </li>
          </ul>
        </section>
      </section>
    </section>
  </middle>
  <back>
    <references anchor="sec-combined-references">
      <name>References</name>
      <references anchor="sec-normative-references">
        <name>Normative References</name>
        <reference anchor="IANA">
          <front>
            <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
            <author fullname="M. Cotton" initials="M." surname="Cotton"/>
            <author fullname="B. Leiba" initials="B." surname="Leiba"/>
            <author fullname="T. Narten" initials="T." surname="Narten"/>
            <date month="June" year="2017"/>
            <abstract>
              <t>Many protocols make use of points of extensibility that use constants to identify various protocol parameters. To ensure that the values in these fields do not have conflicting uses and to promote interoperability, their allocations are often coordinated by a central record keeper. For IETF protocols, that role is filled by the Internet Assigned Numbers Authority (IANA).</t>
              <t>To make assignments in a given registry prudently, guidance describing the conditions under which new values should be assigned, as well as when and how modifications to existing values can be made, is needed. This document defines a framework for the documentation of these guidelines by specification authors, in order to assure that the provided guidance for the IANA Considerations is clear and addresses the various issues that are likely in the operation of a registry.</t>
              <t>This is the third edition of this document; it obsoletes RFC 5226.</t>
            </abstract>
          </front>
          <seriesInfo name="BCP" value="26"/>
          <seriesInfo name="RFC" value="8126"/>
          <seriesInfo name="DOI" value="10.17487/RFC8126"/>
        </reference>
        <reference anchor="HTTP">
          <front>
            <title>HTTP Semantics</title>
            <author fullname="R. Fielding" initials="R." role="editor" surname="Fielding"/>
            <author fullname="M. Nottingham" initials="M." role="editor" surname="Nottingham"/>
            <author fullname="J. Reschke" initials="J." role="editor" surname="Reschke"/>
            <date month="June" year="2022"/>
            <abstract>
              <t>The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document describes the overall architecture of HTTP, establishes common terminology, and defines aspects of the protocol that are shared by all versions. In this definition are core protocol elements, extensibility mechanisms, and the "http" and "https" Uniform Resource Identifier (URI) schemes.</t>
              <t>This document updates RFC 3864 and obsoletes RFCs 2818, 7231, 7232, 7233, 7235, 7538, 7615, 7694, and portions of 7230.</t>
            </abstract>
          </front>
          <seriesInfo name="STD" value="97"/>
          <seriesInfo name="RFC" value="9110"/>
          <seriesInfo name="DOI" value="10.17487/RFC9110"/>
        </reference>
        <reference anchor="PROBLEM">
          <front>
            <title>Problem Details for HTTP APIs</title>
            <author fullname="M. Nottingham" initials="M." surname="Nottingham"/>
            <author fullname="E. Wilde" initials="E." surname="Wilde"/>
            <author fullname="S. Dalal" initials="S." surname="Dalal"/>
            <date month="July" year="2023"/>
            <abstract>
              <t>This document defines a "problem detail" to carry machine-readable details of errors in HTTP response content to avoid the need to define new error response formats for HTTP APIs.</t>
              <t>This document obsoletes RFC 7807.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9457"/>
          <seriesInfo name="DOI" value="10.17487/RFC9457"/>
        </reference>
        <reference anchor="STRUCTURED-FIELDS">
          <front>
            <title>Structured Field Values for HTTP</title>
            <author fullname="M. Nottingham" initials="M." surname="Nottingham"/>
            <author fullname="P-H. Kamp" surname="P-H. Kamp"/>
            <date month="September" year="2024"/>
            <abstract>
              <t>This document describes a set of data types and associated algorithms that are intended to make it easier and safer to define and handle HTTP header and trailer fields, known as "Structured Fields", "Structured Headers", or "Structured Trailers". It is intended for use by specifications of new HTTP fields.</t>
              <t>This document obsoletes RFC 8941.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="9651"/>
          <seriesInfo name="DOI" value="10.17487/RFC9651"/>
        </reference>
        <reference anchor="RFC2119">
          <front>
            <title>Key words for use in RFCs to Indicate Requirement Levels</title>
            <author fullname="S. Bradner" initials="S." surname="Bradner"/>
            <date month="March" year="1997"/>
            <abstract>
              <t>In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements.</t>
            </abstract>
          </front>
          <seriesInfo name="BCP" value="14"/>
          <seriesInfo name="RFC" value="2119"/>
          <seriesInfo name="DOI" value="10.17487/RFC2119"/>
        </reference>
        <reference anchor="RFC8174">
          <front>
            <title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title>
            <author fullname="B. Leiba" initials="B." surname="Leiba"/>
            <date month="May" year="2017"/>
            <abstract>
              <t>RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings.</t>
            </abstract>
          </front>
          <seriesInfo name="BCP" value="14"/>
          <seriesInfo name="RFC" value="8174"/>
          <seriesInfo name="DOI" value="10.17487/RFC8174"/>
        </reference>
      </references>
      <references anchor="sec-informative-references">
        <name>Informative References</name>
        <reference anchor="PRIVACY">
          <front>
            <title>Privacy Considerations for Internet Protocols</title>
            <author fullname="A. Cooper" initials="A." surname="Cooper"/>
            <author fullname="H. Tschofenig" initials="H." surname="Tschofenig"/>
            <author fullname="B. Aboba" initials="B." surname="Aboba"/>
            <author fullname="J. Peterson" initials="J." surname="Peterson"/>
            <author fullname="J. Morris" initials="J." surname="Morris"/>
            <author fullname="M. Hansen" initials="M." surname="Hansen"/>
            <author fullname="R. Smith" initials="R." surname="Smith"/>
            <date month="July" year="2013"/>
            <abstract>
              <t>This document offers guidance for developing privacy considerations for inclusion in protocol specifications. It aims to make designers, implementers, and users of Internet protocols aware of privacy-related design choices. It suggests that whether any individual RFC warrants a specific privacy considerations section will depend on the document's content.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="6973"/>
          <seriesInfo name="DOI" value="10.17487/RFC6973"/>
        </reference>
        <reference anchor="UNIX">
          <front>
            <title>The Single UNIX Specification, Version 2 - 6 Vol Set for UNIX 98</title>
            <author initials="" surname="The Open Group" fullname="The Open Group">
              <organization/>
            </author>
            <date year="1997" month="February"/>
          </front>
        </reference>
        <reference anchor="HPACK">
          <front>
            <title>HPACK: Header Compression for HTTP/2</title>
            <author fullname="R. Peon" initials="R." surname="Peon"/>
            <author fullname="H. Ruellan" initials="H." surname="Ruellan"/>
            <date month="May" year="2015"/>
            <abstract>
              <t>This specification defines HPACK, a compression format for efficiently representing HTTP header fields, to be used in HTTP/2.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="7541"/>
          <seriesInfo name="DOI" value="10.17487/RFC7541"/>
        </reference>
        <reference anchor="HTTP-CACHING">
          <front>
            <title>HTTP Caching</title>
            <author fullname="R. Fielding" initials="R." role="editor" surname="Fielding"/>
            <author fullname="M. Nottingham" initials="M." role="editor" surname="Nottingham"/>
            <author fullname="J. Reschke" initials="J." role="editor" surname="Reschke"/>
            <date month="June" year="2022"/>
            <abstract>
              <t>The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document defines HTTP caches and the associated header fields that control cache behavior or indicate cacheable response messages.</t>
              <t>This document obsoletes RFC 7234.</t>
            </abstract>
          </front>
          <seriesInfo name="STD" value="98"/>
          <seriesInfo name="RFC" value="9111"/>
          <seriesInfo name="DOI" value="10.17487/RFC9111"/>
        </reference>
        <reference anchor="RFC6585">
          <front>
            <title>Additional HTTP Status Codes</title>
            <author fullname="M. Nottingham" initials="M." surname="Nottingham"/>
            <author fullname="R. Fielding" initials="R." surname="Fielding"/>
            <date month="April" year="2012"/>
            <abstract>
              <t>This document specifies additional HyperText Transfer Protocol (HTTP) status codes for a variety of common situations. [STANDARDS-TRACK]</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="6585"/>
          <seriesInfo name="DOI" value="10.17487/RFC6585"/>
        </reference>
        <reference anchor="RFC3339">
          <front>
            <title>Date and Time on the Internet: Timestamps</title>
            <author fullname="G. Klyne" initials="G." surname="Klyne"/>
            <author fullname="C. Newman" initials="C." surname="Newman"/>
            <date month="July" year="2002"/>
            <abstract>
              <t>This document defines a date and time format for use in Internet protocols that is a profile of the ISO 8601 standard for representation of dates and times using the Gregorian calendar.</t>
            </abstract>
          </front>
          <seriesInfo name="RFC" value="3339"/>
          <seriesInfo name="DOI" value="10.17487/RFC3339"/>
        </reference>
      </references>
    </references>
    <?line 733?>

<section anchor="examples">
      <name>Examples</name>
      <section anchor="responses-without-defining-policies">
        <name>Responses without defining policies</name>
        <t>Some servers may not expose the policy limits in the RateLimit-Policy header field. Clients can still use the RateLimit header field to throttle their requests.</t>
        <section anchor="throttling-information-in-responses">
          <name>Throttling information in responses</name>
          <t>The client exhausted its quota for the next 50 seconds.
The limit and time-window is communicated out-of-band.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit: "default";a=0;w=50

{"hello": "world"}
]]></sourcecode>
          <t>Since the field values are not necessarily correlated with
the response status code,
a subsequent request is not required to fail.
The example below shows that the server decided to serve the request
even if the available quota is 0.
Another server, or the same server under other load conditions, could have decided to throttle the request instead.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/456 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit: "default";a=0;w=48

{"still": "successful"}
]]></sourcecode>
        </section>
        <section anchor="use-with-multiple-policies">
          <name>Multiple policies in response</name>
          <t>The server uses two different policies to limit the client's requests:</t>
          <ul spacing="normal">
            <li>
              <t>5000 daily quota units;</t>
            </li>
            <li>
              <t>1000 hourly quota units.</t>
            </li>
          </ul>
          <t>The client consumed 4900 quota units in the first 14 hours.</t>
          <t>Despite the next hourly limit of 1000 quota units,
the closest limit to reach is the daily one.</t>
          <t>The server then exposes the RateLimit header fields to
inform the client that:</t>
          <ul spacing="normal">
            <li>
              <t>it has only 100 quota units left in the daily quota and the window will end in 10 hours;</t>
            </li>
          </ul>
          <t>The server MAY choose to omit returning the hourly policy as it uses the same quota units as the daily policy and the daily policy is the one that is closest to being exhausted.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit: "dayLimit";a=100;w=36000

{"hello": "world"}
]]></sourcecode>
        </section>
        <section anchor="use-for-limiting-concurrency">
          <name>Use for limiting concurrency</name>
          <t>RateLimit header fields may be used to limit concurrency,
advertising limits that are lower than the usual ones
in case of saturation, thus increasing availability.</t>
          <t>The server adopted a basic policy of 100 quota units per minute,
and in case of resource exhaustion adapts the returned values
reducing both available quota and effective window values.</t>
          <t>After 2 seconds the client consumed 40 quota units</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit-Policy: "basic";q=100;w=60
RateLimit: "basic";a=60;w=58

{"elapsed": 2, "issued": 40}
]]></sourcecode>
          <t>At the subsequent request - due to resource exhaustion -
the server advertises only <tt>r=20</tt>.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit-Policy: "basic";q=100;w=60
RateLimit: "basic";a=20;w=56

{"elapsed": 4, "issued": 41}
]]></sourcecode>
        </section>
        <section anchor="use-in-throttled-responses">
          <name>Use in throttled responses</name>
          <t>A client exhausted its quota and the server throttles it
sending Retry-After.</t>
          <t>In this example, the values of Retry-After and RateLimit header field reference the same moment,
but this is not a requirement.</t>
          <t>The 429 (Too Many Request) HTTP status code is just used as an example.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 429 Too Many Requests
Content-Type: application/problem+json
Date: Mon, 05 Aug 2019 09:27:00 GMT
Retry-After: Mon, 05 Aug 2019 09:27:05 GMT
RateLimit: "default";a=0;w=5

{
"type": "https://iana.org/assignments/http-problem-types#quota-exceeded"
"title": "Too Many Requests",
"status": 429,
"violated-policies": ["default"]
}
]]></sourcecode>
        </section>
      </section>
      <section anchor="responses-with-defined-policies">
        <name>Responses with defined policies</name>
        <section anchor="throttling-window-specified-via-parameter">
          <name>Throttling window specified via parameter</name>
          <t>The client has 99 quota units left for the next 50 seconds.
The time window is communicated by the <tt>w</tt> parameter, so we know the throughput is 100 quota units per minute.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit: "fixedwindow";a=99;w=50
RateLimit-Policy: "fixedwindow";q=100;w=60
{"hello": "world"}
]]></sourcecode>
        </section>
        <section anchor="dynamic-limits-with-parameterized-windows">
          <name>Dynamic limits with parameterized windows</name>
          <t>The policy conveyed by the RateLimit-Policy header field states that
the server accepts 100 quota units per minute.</t>
          <t>To avoid resource exhaustion, the server artificially lowers
the actual limits returned in the throttling headers.</t>
          <t>The RateLimit header field then advertises
only 9 quota units for the next 50 seconds to slow down the client.</t>
          <t>Note that the server could have lowered even the other
values in the RateLimit header field: this specification
does not mandate any relation between the field values
contained in subsequent responses.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 Ok
Content-Type: application/json
RateLimit-Policy: "dynamic";q=100;w=60
RateLimit: "dynamic";a=9;w=50


{
  "status": 200,
  "detail": "Just slow down without waiting."
}
]]></sourcecode>
        </section>
        <section anchor="dynamic-limits-for-pushing-back-and-slowing-down">
          <name>Dynamic limits for pushing back and slowing down</name>
          <t>Continuing the previous example, let's say the client waits 10 seconds and
performs a new request which, due to resource exhaustion, the server rejects
and pushes back, advertising <tt>r=0</tt> for the next 20 seconds.</t>
          <t>The server advertises a smaller window with a lower limit to slow
down the client for the rest of its original window after the 20 seconds elapse.</t>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 429 Too Many Requests
Content-Type: application/json
RateLimit-Policy: "dynamic";q=15;w=20
RateLimit: "dynamic";a=0;w=20

{
  "status": 429,
  "detail": "Wait 20 seconds, then slow down!"
}
]]></sourcecode>
        </section>
      </section>
      <section anchor="dynamic-limits-for-pushing-back-with-retry-after-and-slow-down">
        <name>Dynamic limits for pushing back with Retry-After and slow down</name>
        <t>Alternatively, given the same context where the previous example starts, we
can convey the same information to the client via Retry-After, with
the advantage that the server can now specify the policy's nominal limit and
window that will apply after the retry time, e.g. assuming the resource exhaustion
is likely to be gone by then, so the advertised policy does not need to be
adjusted, yet we managed to stop requests for a while and slow down the rest of
the current window.</t>
        <t>Request:</t>
        <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
        <t>Response:</t>
        <sourcecode type="http-message"><![CDATA[
HTTP/1.1 429 Too Many Requests
Content-Type: application/json
Retry-After: 20
RateLimit-Policy: "dynamic";q=100;w=60
RateLimit: "dynamic";a=15;w=40

{
  "status": 429,
  "detail": "Wait 20 seconds, then slow down!"
}
]]></sourcecode>
        <t>Note that in this last response the client is expected to honor
Retry-After and perform no requests for the specified amount of
time, whereas the previous example would not force the client to stop
requests before the end of the effective window, as it would still be free to
query again the server even if it is likely to have the request rejected.</t>
        <section anchor="use-with-multiple-windows">
          <name>Use with multiple windows</name>
          <t>This is a standardized way of describing the policy
detailed in <xref target="use-with-multiple-policies"/>:</t>
          <ul spacing="normal">
            <li>
              <t>5000 daily quota units;</t>
            </li>
            <li>
              <t>1000 hourly quota units.</t>
            </li>
          </ul>
          <t>The client consumed 4900 quota units in the first 14 hours.</t>
          <t>Despite the next hourly limit of 1000 quota units, the closest limit
to reach is the daily one.</t>
          <t>The server then exposes the RateLimit header fields to inform the client that:</t>
          <ul spacing="normal">
            <li>
              <t>it has only 100 quota units left;</t>
            </li>
            <li>
              <t>the effective window is 10 hours;</t>
            </li>
            <li>
              <t>the expiring-limit is 5000.</t>
            </li>
          </ul>
          <t>Request:</t>
          <sourcecode type="http-message"><![CDATA[
GET /items/123 HTTP/1.1
Host: api.example

]]></sourcecode>
          <t>Response:</t>
          <sourcecode type="http-message"><![CDATA[
HTTP/1.1 200 OK
Content-Type: application/json
RateLimit-Policy: "hour";q=1000;w=3600, "day";q=5000;w=86400
RateLimit: "day";a=100;w=36000

{"hello": "world"}
]]></sourcecode>
        </section>
      </section>
    </section>
    <section numbered="false" removeInRFC="true" anchor="faq">
      <name>FAQ</name>
      <ol spacing="normal" type="1"><li>
          <t>Why defining standard fields for throttling?  </t>
          <t>
To simplify enforcement of throttling policies and enable clients to constraint their requests to avoid being throttled.</t>
        </li>
        <li>
          <t>Can I use RateLimit header fields in throttled responses (e.g. with status code 429)?  </t>
          <t>
Yes, you can.</t>
        </li>
        <li>
          <t>Are those specs tied to RFC 6585?  </t>
          <t>
No. <xref target="RFC6585"/> defines the <tt>429</tt> status code and we use it just as an example of a throttled request,
that could instead use even <tt>403</tt> or whatever status code.</t>
        </li>
        <li>
          <t>Why is the partition key necessary?  </t>
          <t>
Without a partition key, a server can effectively only have one scope (aka partition), which is impractical for most services, or it needs to communicate the scopes out-of-band.
This prevents the development of generic connector code that can be used to prevent requests from being throttled.
Many APIs rely on API keys, user identity or client identity to allocate quota.
As soon as a single client processes requests for more than one partition, the client needs to know the corresponding partition key to properly track requests against allocated quota.</t>
        </li>
        <li>
          <t>Why using delay-seconds instead of a UNIX Timestamp?
Why not using subsecond precision?  </t>
          <t>
Using delay-seconds aligns with Retry-After, which is returned in similar contexts,
e.g. on 429 responses.  </t>
          <t>
Timestamps require a clock synchronization protocol
(see <xref section="5.6.7" sectionFormat="of" target="HTTP"/>).
This may be problematic (e.g. clock adjustment, clock skew, failure of hardcoded clock synchronization servers,
IoT devices, etc.).
Moreover timestamps may not be monotonically increasing due to clock adjustment.
See <eref target="https://community.ntppool.org/t/another-ntp-client-failure-story/1014/">Another NTP client failure story</eref>  </t>
          <t>
We did not use subsecond precision because:  </t>
          <ul spacing="normal">
            <li>
              <t>that is more subject to system clock correction
like the one implemented via the adjtimex() Linux system call;</t>
            </li>
            <li>
              <t>response-time latency may not make it worth. A brief discussion on the subject is
on the <eref target="https://lists.w3.org/Archives/Public/ietf-http-wg/2019JulSep/0202.html">httpwg ml</eref></t>
            </li>
            <li>
              <t>almost all rate-limit headers implementations do not use it.</t>
            </li>
          </ul>
        </li>
        <li>
          <t>Shouldn't I limit concurrency instead of request rate?  </t>
          <t>
You can use this specification to limit concurrency
at the HTTP level (see {#use-for-limiting-concurrency})
and help clients to shape their requests avoiding being throttled out.  </t>
          <t>
A problematic way to limit concurrency is connection dropping,
especially when connections are multiplexed (e.g. HTTP/2)
because this results in client requests not being handled,
which is something we want to avoid.  </t>
          <t>
A semantic way to limit concurrency is to return 503 + Retry-After
in case of resource saturation (e.g. thrashing, connection queues too long,
Service Level Objectives not meet, etc.).
Saturation conditions can be either dynamic or static: all this is out of
the scope for the current document.</t>
        </li>
        <li>
          <t>Do a positive value of remaining parameter imply any service guarantee for my
future requests to be served?  </t>
          <t>
No. FAQ integrated in <xref target="ratelimit-available-quota"/>.</t>
        </li>
        <li>
          <t>Is the quota-policy definition too complex?  </t>
          <t>
You can always return the simplest form</t>
        </li>
      </ol>
      <artwork><![CDATA[
RateLimit:"default";a=50;w=60
]]></artwork>
      <t>The policy key clearly connects the current usage status of a policy to the defined limits.
   So for the following field:</t>
      <artwork><![CDATA[
RateLimit-Policy: "sliding";q=100;w=60;burst=1000
RateLimit-Policy: "fixed";q=5000;w=3600;burst=0
RateLimit: "sliding";a=50;w=44
]]></artwork>
      <t>the value "sliding" identifies the policy being reported.</t>
      <ol spacing="normal" type="1"><li>
          <t>Can intermediaries alter RateLimit header fields?  </t>
          <t>
Generally, they should not because it might result in requests not being handled.
 There are reasonable use cases for intermediaries mangling RateLimit header fields though,
 e.g. when they enforce stricter quota-policies,
 or when they are an active component of the service.
 In those cases we will consider them as part of the originating infrastructure.</t>
        </li>
        <li>
          <t>Why the <tt>w</tt> parameter is just informative?
Could it be used by a client to determine the request rate?  </t>
          <t>
A non-informative <tt>w</tt> parameter might be fine in an environment
where clients and servers are tightly coupled. Conveying policies
with this detail on a large scale would be very complex and implementations
would likely be not interoperable. We thus decided to leave <tt>w</tt> as
an informational parameter and only rely on the limit, remaining and reset parameters
for defining the throttling
behavior.</t>
        </li>
        <li>
          <t>Can I use RateLimit fields in trailers?  </t>
          <t>
No. Servers usually establish whether the request is within quota before creating a response, so RateLimit field values are generally available when the header section is generated. Using trailers would complicate client implementations and intermediaries might drop them. Servers therefore MUST NOT generate RateLimit fields in a trailer section, and recipients MUST ignore them if received there.</t>
        </li>
      </ol>
    </section>
    <section numbered="false" removeInRFC="true" anchor="ratelimit-header-fields-currently-used-on-the-web">
      <name>RateLimit header fields currently used on the web</name>
      <t>Commonly used header field names are:</t>
      <ul spacing="normal">
        <li>
          <t><tt>X-RateLimit-Limit</tt>,
<tt>X-RateLimit-Remaining</tt>,
<tt>X-RateLimit-Reset</tt>;</t>
        </li>
      </ul>
      <t>There are variants too, where the window is specified
in the header field name, e.g.:</t>
      <ul spacing="normal">
        <li>
          <t><tt>x-ratelimit-limit-minute</tt>, <tt>x-ratelimit-limit-hour</tt>, <tt>x-ratelimit-limit-day</tt></t>
        </li>
        <li>
          <t><tt>x-ratelimit-remaining-minute</tt>, <tt>x-ratelimit-remaining-hour</tt>, <tt>x-ratelimit-remaining-day</tt></t>
        </li>
      </ul>
      <t>Here are some interoperability issues:</t>
      <ul spacing="normal">
        <li>
          <t><tt>X-RateLimit-Remaining</tt> references different values, depending on the implementation:  </t>
          <ul spacing="normal">
            <li>
              <t>seconds remaining to the window expiration</t>
            </li>
            <li>
              <t>milliseconds remaining to the window expiration</t>
            </li>
            <li>
              <t>seconds since UTC, in UNIX Timestamp <xref target="UNIX"/></t>
            </li>
            <li>
              <t>a datetime, either <tt>IMF-fixdate</tt> <xref target="HTTP"/> or <xref target="RFC3339"/></t>
            </li>
          </ul>
        </li>
        <li>
          <t>different headers, with the same semantic, are used by different implementers:  </t>
          <ul spacing="normal">
            <li>
              <t>X-RateLimit-Limit and X-Rate-Limit-Limit</t>
            </li>
            <li>
              <t>X-RateLimit-Remaining and X-Rate-Limit-Remaining</t>
            </li>
            <li>
              <t>X-RateLimit-Reset and X-Rate-Limit-Reset</t>
            </li>
          </ul>
        </li>
      </ul>
      <t>The semantic of RateLimit depends on the windowing algorithm.
A sliding window policy for example, may result in having a remaining parameter value related to the ratio between the current and the maximum throughput.
e.g.</t>
      <artwork><![CDATA[
RateLimit-Policy: "sliding";q=12;w=1
; using 50% of throughput, that is 6 units/s
RateLimit: "sliding";q=12;a=6;w=1

]]></artwork>
      <t>If this is the case, the optimal solution is to achieve</t>
      <artwork><![CDATA[
RateLimit-Policy: "sliding";q=12;w=1
; using 100% of throughput, that is 12 units/s
RateLimit: "sliding";q=12;a=1;w=1
]]></artwork>
      <t>At this point you should stop increasing your request rate.</t>
    </section>
    <section numbered="false" anchor="acknowledgements">
      <name>Acknowledgements</name>
      <t>Thanks to Willi Schoenborn, Alejandro Martinez Ruiz, Alessandro Ranellucci,
Amos Jeffries,
Martin Thomson,
Erik Wilde and Mark Nottingham for being the initial contributors
of these specifications.
Kudos to the first community implementers:
Aapo Talvensaari,
Nathan Friedly
and Sanyam Dogra.</t>
      <t>In addition to the people above, this document owes a lot to the extensive discussion in the HTTPAPI workgroup, including
Rich Salz,
and Julian Reschke.</t>
    </section>
    <section numbered="false" removeInRFC="true" anchor="changes">
      <name>Changes</name>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-08">
        <name>Since draft-ietf-httpapi-ratelimit-headers-08</name>
        <ul spacing="normal">
          <li>
            <t>Added Problem Types</t>
          </li>
          <li>
            <t>Clarified when to use RateLimit-Policy vs RateLimit fields</t>
          </li>
        </ul>
      </section>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-07">
        <name>Since draft-ietf-httpapi-ratelimit-headers-07</name>
        <ul spacing="normal">
          <li>
            <t>Refactored both fields to lists of Items that identify policy and use parameters</t>
          </li>
          <li>
            <t>Added quota unit parameter</t>
          </li>
          <li>
            <t>Added partition key parameter</t>
          </li>
        </ul>
      </section>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-03">
        <name>Since draft-ietf-httpapi-ratelimit-headers-03</name>
        <ul spacing="normal">
          <li>
            <t>Split policy informatiom in RateLimit-Policy #81</t>
          </li>
        </ul>
      </section>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-02">
        <name>Since draft-ietf-httpapi-ratelimit-headers-02</name>
        <ul spacing="normal">
          <li>
            <t>Address throttling scope #83</t>
          </li>
        </ul>
      </section>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-01">
        <name>Since draft-ietf-httpapi-ratelimit-headers-01</name>
        <ul spacing="normal">
          <li>
            <t>Update IANA considerations #60</t>
          </li>
          <li>
            <t>Use Structured fields #58</t>
          </li>
          <li>
            <t>Reorganize document #67</t>
          </li>
        </ul>
      </section>
      <section numbered="false" removeInRFC="true" anchor="since-draft-ietf-httpapi-ratelimit-headers-00">
        <name>Since draft-ietf-httpapi-ratelimit-headers-00</name>
        <ul spacing="normal">
          <li>
            <t>Use I-D.httpbis-semantics, which includes referencing delay-seconds
instead of delta-seconds. #5</t>
          </li>
        </ul>
      </section>
    </section>
  </back>
  <!-- ##markdown-source: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-->

</rfc>
