Internet-Draft | RateLimit Header Fields for HTTP | May 2020 |
Polli & Martinez | Expires 27 November 2020 | [Page] |
This document defines the RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset header fields for HTTP, thus allowing servers to publish current request quotas and clients to shape their request policy and avoid being throttled out.¶
RFC EDITOR: please remove this section before publication¶
Discussion of this draft takes place on the HTTP working group mailing list (ietf-http-wg@w3.org), which is archived at https://lists.w3.org/Archives/Public/ietf-http-wg/.¶
The source code and issues list for this draft can be found at https://github.com/ioggstream/draft-polli-ratelimit-headers.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 27 November 2020.¶
Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
The widespreading of HTTP as a distributed computation protocol requires an explicit way of communicating service status and usage quotas.¶
This was partially addressed with the Retry-After
header field
defined in [RFC7231] to be returned in 429 Too Many Requests
or
503 Service Unavailable
responses.¶
Still, there is not a standard way to communicate service quotas so that the client can throttle its requests and prevent 4xx or 5xx responses.¶
Servers use quota mechanisms to avoid systems overload, to ensure an equitable distribution of computational resources or to enforce other policies - eg. monetization.¶
A basic quota mechanism limits the number of acceptable requests in a given time window, eg. 10 requests per second.¶
When quota is exceeded, servers usually do not serve the request
replying instead with a 4xx
HTTP status code (eg. 429 or 403)
or adopt more aggressive policies like dropping connections.¶
Quotas may be enforced on different basis (eg. per user, per IP, per geographic area, ..) and at different levels. For example, an user may be allowed to issue:¶
Moreover system metrics, statistics and heuristics can be used to implement more complex policies, where the number of acceptable request and the time window are computed dynamically.¶
To help clients throttling their requests, servers may expose the counters used to evaluate quota policies via HTTP header fields.¶
Those response headers may be added by HTTP intermediaries such as API gateways and reverse proxies.¶
On the web we can find many different rate-limit headers, usually containing the number of allowed requests in a given time window, and when the window is reset.¶
The common choice is to return three headers containing:¶
A major interoperability issue in throttling is the lack of standard headers, because:¶
Client applications interfacing with different servers may thus need to process different headers, or the very same application interface that sits behind different reverse proxies may reply with different throttling headers.¶
This proposal defines syntax and semantics for the following header fields:¶
RateLimit-Limit
: containing the requests quota in the time window;¶
RateLimit-Remaining
: containing the remaining requests quota in the current window;¶
RateLimit-Reset
: containing the time remaining in the current window, specified in seconds.¶
The behavior of RateLimit-Reset
is compatible with the delta-seconds
notation of Retry-After
.¶
The header fields definition allows to describe complex policies, including the ones using multiple and variable time windows and dynamic quotas, or implementing concurrency limits.¶
The goals of this proposal are:¶
The goals do not include:¶
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 ([RFC2119] and [RFC8174]) when, and only when, they appear in all capitals, as shown here.¶
This document uses the Augmented BNF defined in [RFC5234] and updated by [RFC7405] along with the "#rule" extension defined in Section 7 of [RFC7230].¶
The term Origin is to be interpreted as described in [RFC6454] section 7.¶
The "delta-seconds" rule is defined in [RFC7234] section 1.2.1.¶
Rate limit policies limit the number of acceptable requests in a given time window.¶
A time window is expressed in seconds, using the following syntax:¶
time-window = delta-seconds¶
Subsecond precision is not supported.¶
The request-quota is a value associated to the maximum number of requests
that the server is willing to accept
from one or more clients
on a given basis (originating IP, authenticated user, geographical, ..)
during a time-window
as defined in Section 2.1.¶
The request-quota
is expressed in quota-units
and has the following syntax:¶
request-quota = quota-units quota-units = 1*DIGIT¶
The request-quota
SHOULD match the maximum number of acceptable requests.¶
The request-quota
MAY differ from the total number of acceptable requests
when weight mechanisms, bursts, or other server policies are implemented.¶
If the request-quota
does not match the maximum number of acceptable requests
the relation with that SHOULD be communicated out-of-band.¶
Example: A server could¶
so that we have the following counters¶
GET /books/123 ; request-quota=4, remaining: 3, status=200 GET /books?author=Camilleri ; request-quota=4, remaining: 1, status=200 GET /books?author=Eco ; request-quota=4, remaining: 0, status=429¶
This specification allows describing a quota policy with the following syntax:¶
quota-policy = request-quota; "w" "=" time-window *( OWS ";" OWS quota-comment) quota-comment = token "=" (token / quoted-string)¶
quota-policy parameters like w
and
quota-comment tokens MUST NOT occur multiple times within the same quota-policy.¶
An example policy of 100 quota-units per minute.¶
100;w=60¶
Two examples of providing further details via custom parameters
in quota-comments
.¶
100;w=60;comment="fixed window" 12;w=1;burst=1000;policy="leaky bucket"¶
The following RateLimit
response header fields are defined¶
The RateLimit-Limit
response header field indicates
the request-quota
associated to the client
in the current time-window
.¶
If the client exceeds that limit, it MAY not be served.¶
The header value is¶
RateLimit-Limit = expiring-limit [, 1#quota-policy ] expiring-limit = request-quota¶
The expiring-limit
value MUST be set to the request-quota
that is closer to reach its limit.¶
The quota-policy
is defined in Section 2.3, and its values are informative.¶
RateLimit-Limit: 100¶
A time-window
associated to expiring-limit
can be communicated
via an optional quota-policy
value, like shown in the following example¶
RateLimit-Limit: 100, 100;w=10¶
If the expiring-limit
is not associated to a time-window
, the time-window
MUST either be:¶
RateLimit-Reset
at the moment of the reset, or¶
Policies using multiple quota limits MAY be returned using multiple
quota-policy
items, like shown in the following two examples:¶
RateLimit-Limit: 10, 10;w=1, 50;w=60, 1000;w=3600, 5000;w=86400 RateLimit-Limit: 10, 10;w=1;burst=1000, 1000;w=3600¶
This header MUST NOT occur multiple times.¶
The RateLimit-Remaining
response header field indicates the remaining quota-units
defined in Section 2.2
associated to the client.¶
The header value is¶
RateLimit-Remaining = quota-units¶
This header MUST NOT occur multiple times.¶
Clients MUST NOT assume that a positive RateLimit-Remaining
value is
a guarantee of being served.¶
A low RateLimit-Remaining
value is like a yellow traffic-light: the red light
may arrive suddenly.¶
One example of RateLimit-Remaining
use is below.¶
RateLimit-Remaining: 50¶
The RateLimit-Reset
response header field indicates either¶
The header value is¶
RateLimit-Reset = delta-seconds¶
The delta-seconds format is used because:¶
This header MUST NOT occur multiple times.¶
An example of RateLimit-Reset
use is below.¶
RateLimit-Reset: 50¶
The client MUST NOT assume that all its request-quota
will be restored
after the moment referenced by RateLimit-Reset
.
The server MAY arbitrarily alter the RateLimit-Reset
value between subsequent requests
eg. in case of resource saturation or to implement sliding window policies.¶
A server MAY use one or more RateLimit
response header fields
defined in this document to communicate its quota policies.¶
The returned values refers to the metrics used to evaluate if the current request respects the quota policy and MAY not apply to subsequent requests.¶
Example: a successful response with the following header fields¶
RateLimit-Limit: 10 RateLimit-Remaining: 1 RateLimit-Reset: 7¶
does not guarantee that the next request will be successful. Server metrics may be subject to other conditions like the one shown in the example from Section 2.2.¶
A server MAY return RateLimit
response header fields independently
of the response status code. This includes throttled responses.¶
If a response contains both the Retry-After
and the RateLimit-Reset
header fields,
the value of RateLimit-Reset
SHOULD reference the same point in time as
Retry-After
.¶
When using a policy involving more than one time-window
,
the server MUST reply with the RateLimit
headers related to the window
with the lower RateLimit-Remaining
values.¶
Under certain conditions, a server MAY artificially lower RateLimit
field values between subsequent requests,
eg. to respond to Denial of Service attacks or in case of resource saturation.¶
A client MUST process the received RateLimit
headers.¶
A client MUST validate the values received in the RateLimit
headers before using them
and check if there are significant discrepancies
with the expected ones.
This includes a RateLimit-Reset
moment too far in the future or a request-quota
too high.¶
Malformed RateLimit
headers MAY be ignored.¶
A client SHOULD NOT exceed the quota-units
expressed in RateLimit-Remaining
before the time-window
expressed
in RateLimit-Reset
.¶
A client MAY still probe the server if the RateLimit-Reset
is considered too high.¶
The value of RateLimit-Reset
is generated at response time:
a client aware of a significant network latency MAY behave accordingly
and use other informations (eg. the Date
response header, or otherwise gathered metrics) to better
estimate the RateLimit-Reset
moment intended by the server.¶
The quota-policy
values and comments provided in RateLimit-Limit
are informative
and MAY be ignored.¶
If a response contains both the RateLimit-Reset
and Retry-After
header fields,
the Retry-After
header field MUST take precedence and
the RateLimit-Reset
header field MAY be ignored.¶
The client exhausted its request-quota for the next 50 seconds.
The time-window
is communicated out-of-band or inferred by the header values.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 100 Ratelimit-Remaining: 0 Ratelimit-Reset: 50 {"hello": "world"}¶
The server uses two custom headers,
namely acme-RateLimit-DayLimit
and acme-RateLimit-HourLimit
to expose the following policy:¶
The client consumed 4900 quota-units in the first 14 hours.¶
Despite the next hourly limit of 1000 quota-units, the closest limit to reach is the daily one.¶
The server then exposes the RateLimit-*
headers to
inform the client that:¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json acme-RateLimit-DayLimit: 5000 acme-RateLimit-HourLimit: 1000 RateLimit-Limit: 5000 RateLimit-Remaining: 100 RateLimit-Reset: 36000 {"hello": "world"}¶
Throttling headers may be used to limit concurrency, advertising limits that are lower than the usual ones in case of saturation, thus increasing availability.¶
The server adopted a basic policy of 100 quota-units
per minute,
and in case of resource exhaustion adapts the returned values
reducing both RateLimit-Limit
and RateLimit-Remaining
.¶
After 2 seconds the client consumed 40 quota-units¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 100 RateLimit-Remaining: 60 RateLimit-Reset: 58 {"elapsed": 2, "issued": 40}¶
At the subsequent request - due to resource exhaustion -
the server advertises only RateLimit-Remaining: 20
.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 100 RateLimit-Remaining: 20 RateLimit-Reset: 56 {"elapsed": 4, "issued": 41}¶
A client exhausted its quota and the server throttles the request
sending the Retry-After
response header field.¶
In this example, the values of Retry-After
and RateLimit-Reset
reference the same moment,
but this is not a requirement.¶
The 429 Too Many Requests
HTTP status code is just used as an example.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 429 Too Many Requests Content-Type: application/json Date: Mon, 05 Aug 2019 09:27:00 GMT Retry-After: Mon, 05 Aug 2019 09:27:05 GMT RateLimit-Reset: 5 RateLimit-Limit: 100 Ratelimit-Remaining: 0 { "title": "Too Many Requests", "status": 429, "detail": "You have exceeded your quota" }¶
The client has 99 quota-units
left for the next 50 seconds.
The time-window
is communicated by the w
parameter, so we know the throughput is 100 quota-units
per minute.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 100, 100;w=60 Ratelimit-Remaining: 99 Ratelimit-Reset: 50 {"hello": "world"}¶
The policy conveyed by RateLimit-Limit
states that
the server accepts 100 quota-units per minute.¶
To avoid resource exhaustion, the server artificially lowers the actual limits returned in the throttling headers.¶
The RateLimit-Remaining
then advertises
only 9 quota-units for the next 50 seconds to slow down the client.¶
Note that the server could have lowered even the other
values in RateLimit-Limit
: this specification
does not mandate any relation between the field values
contained in subsequent responses.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 10, 100;w=60 Ratelimit-Remaining: 9 Ratelimit-Reset: 50 { "status": 200, "detail": "Just slow down without waiting." }¶
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 RateLimit-Remaining: 0
for the next 20 seconds.¶
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.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 429 Too Many Requests Content-Type: application/json RateLimit-Limit: 0, 15;w=20 Ratelimit-Remaining: 0 Ratelimit-Reset: 20 { "status": 429, "detail": "Wait 20 seconds, then slow down!" }¶
Alternatively, given the same context where the previous example starts, we can convey the same information to the client via the Retry-After header, with the advantage that the server can now specify the policy's nominal limit and window that will apply after the reset, ie. 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.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 429 Too Many Requests Content-Type: application/json Retry-After: 20 RateLimit-Limit: 15, 100;w=60 Ratelimit-Remaining: 15 Ratelimit-Reset: 40 { "status": 429, "detail": "Wait 20 seconds, then slow down!" }¶
Note that in this last response the client is expected to honor the
Retry-After
header and perform no requests for the specified amount of
time, whereas the previous example would not force the client to stop
requests before the reset time is elapsed, as it would still be free to
query again the server even if it is likely to have the request rejected.¶
The server does not expose RateLimit-Remaining
values, but
resets the limit counter every second.¶
It communicates to the client the limit of 10 quota-units per second
always returning the couple RateLimit-Limit
and RateLimit-Reset
.¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 10 Ratelimit-Reset: 1 {"first": "request"}¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 Ok Content-Type: application/json RateLimit-Limit: 10 Ratelimit-Reset: 1 {"second": "request"}¶
This is a standardized way of describing the policy detailed in Section 6.1.2:¶
The client consumed 4900 quota-units in the first 14 hours.¶
Despite the next hourly limit of 1000 quota-units, the closest limit to reach is the daily one.¶
The server then exposes the RateLimit
headers to
inform the client that:¶
Request:¶
GET /items/123¶
Response:¶
HTTP/1.1 200 OK Content-Type: application/json RateLimit-Limit: 5000, 1000;w=3600, 5000;w=86400 RateLimit-Remaining: 100 RateLimit-Reset: 36000 {"hello": "world"}¶
This specification does not prevent clients to make over-quota requests.¶
Servers should always implement mechanisms to prevent resource exhaustion.¶
Servers should not disclose operational capacity informations that can be used to saturate its resources.¶
While this specification does not mandate whether non 2xx responses consume quota, if 401 and 403 responses count on quota a malicious client could probe the endpoint to get traffic informations of another user.¶
RateLimit-*
headers convey hints from the server
to the clients in order to avoid being throttled out.¶
Clients MUST NOT consider the quota-units
returned in RateLimit-Remaining
as a service level agreement.¶
In case of resource saturation, the server MAY artificially lower the returned values or not serve the request anyway.¶
Consider that request-quota
may not be restored after the moment referenced by RateLimit-Reset
,
and the RateLimit-Reset
value should not be considered fixed nor constant.¶
Subsequent requests may return an higher RateLimit-Reset
value to limit
concurrency or implement dynamic or adaptive throttling policies.¶
When returning RateLimit-Reset
you must be aware that many throttled
clients may come back at the very moment specified.¶
This is true for Retry-After
too.¶
For example, if the quota resets every day at 18:00:00
and your server returns the RateLimit-Reset
accordingly¶
Date: Tue, 15 Nov 1994 08:00:00 GMT RateLimit-Reset: 36000¶
there's a high probability that all clients will show up at 18:00:00
.¶
This could be mitigated adding some jitter to the field-value.¶
RateLimit
header fields may assume unexpected values by chance or purpose.
For example, an excessively high RateLimit-Remaining
value may be:¶
or an high RateLimit-Reset
value could inhibit clients to contact
the server.¶
Clients MUST validate the received values to mitigate those risks.¶
This section registers the RateLimit-Limit
header field in the "Permanent Message
Header Field Names" registry ([RFC3864]).¶
Header field name: RateLimit-Limit
¶
Applicable protocol: http¶
Status: standard¶
Author/Change controller: IETF¶
Specification document(s): Section 3.1 of this document¶
This section registers the RateLimit-Remaining
header field in the "Permanent Message
Header Field Names" registry ([RFC3864]).¶
Header field name: RateLimit-Remaining
¶
Applicable protocol: http¶
Status: standard¶
Author/Change controller: IETF¶
Specification document(s): Section 3.2 of this document¶
This section registers the RateLimit-Reset
header field in the "Permanent Message
Header Field Names" registry ([RFC3864]).¶
Header field name: RateLimit-Reset
¶
Applicable protocol: http¶
Status: standard¶
Author/Change controller: IETF¶
Specification document(s): Section 3.3 of this document¶
RFC EDITOR PLEASE DELETE THIS SECTION.¶
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 implementors: Aapo Talvensaari, Nathan Friedly and Sanyam Dogra.¶
RFC EDITOR PLEASE DELETE THIS SECTION.¶
Commonly used header field names are:¶
X-RateLimit-Limit
, X-RateLimit-Remaining
, X-RateLimit-Reset
;¶
X-Rate-Limit-Limit
, X-Rate-Limit-Remaining
, X-Rate-Limit-Reset
.¶
There are variants too, where the window is specified in the header field name, eg:¶
x-ratelimit-limit-minute
, x-ratelimit-limit-hour
, x-ratelimit-limit-day
¶
x-ratelimit-remaining-minute
, x-ratelimit-remaining-hour
, x-ratelimit-remaining-day
¶
Here are some interoperability issues:¶
X-RateLimit-Remaining
references different values, depending on the implementation:¶
different headers, with the same semantic, are used by different implementers:¶
The semantic of RateLimit-Remaining depends on the windowing algorithm. A sliding window policy for example may result in having a ratelimit-remaining value related to the ratio between the current and the maximum throughput. Eg.¶
RateLimit-Limit: 12, 12;w=1 RateLimit-Remaining: 6 ; using 50% of throughput, that is 6 units/s RateLimit-Reset: 1¶
If this is the case, the optimal solution is to achieve¶
RateLimit-Limit: 12, 12;w=1 RateLimit-Remaining: 1 ; using 100% of throughput, that is 12 units/s RateLimit-Reset: 1¶
At this point you should stop increasing your request rate.¶
Why defining standard headers for throttling?¶
To simplify enforcement of throttling policies.¶
Can I use RateLimit-* in throttled responses (eg with status code 429)?¶
Yes, you can.¶
Are those specs tied to RFC 6585?¶
No. [RFC6585] defines the 429
status code and we use it just as an example of a throttled request,
that could instead use even 403 or whatever status code.¶
Why don't pass the trottling scope as a parameter?¶
I'm open to suggestions. File an issue if you think it's worth ;).¶
Why using delta-seconds instead of a UNIX Timestamp? Why not using subsecond precision?¶
Using delta-seconds aligns with Retry-After
, which is returned in similar contexts,
eg on 429 responses.¶
delta-seconds as defined in [RFC7234] section 1.2.1 clarifies some parsing rules too.¶
Timestamps require a clock synchronization protocol (see [RFC7231] section 4.1.1.1). This may be problematic (eg. clock adjustment, clock skew, failure of hardcoded clock synchronization servers, IoT devices, ..). Moreover timestamps may not be monotonically increasing due to clock adjustment. See Another NTP client failure story¶
We did not use subsecond precision because:¶
Why not support multiple quota remaining?¶
While this might be of some value, my experience suggests that overly-complex quota implementations results in lower effectiveness of this policy. This spec allows the client to easily focusing on RateLimit-Remaining and RateLimit-Reset.¶
Shouldn't I limit concurrency instead of request rate?¶
You can do both. The goal of this spec is to provide guidance for clients in shaping their requests without being throttled out.¶
Limiting concurrency results in unserviced client requests, which is something we want to avoid.¶
A standard way to limit concurrency is to return 503 + Retry-After in case of resource saturation (eg. thrashing, connection queues too long, Service Level Objectives not meet, ..).¶
Availability can be improved by dynamically lowering the values returned by
the RateLimit-*
headers to slow down clients, and Retry-After
can be used to
push them back.¶
Saturation conditions can be either dynamic or static: all this is out of the scope for the current document.¶
Do a positive value of RateLimit-Remaining
imply any service guarantee for my
future requests to be served?¶
No. The returned values were used to decide whether to serve or not the current request and do not imply any guarantee that future requests will be successful.¶
Instead they help to understand when future requests
will probably be throttled. A low value for RateLimit-Remaining
should be interpreted as a yellow traffic-light for either
the number of requests issued in the time-window
or the request throughput.¶
Is the quota-policy definition Section 2.3 too complex?¶
You can always return the simplest form of the 3 headers¶
RateLimit-Limit: 100 RateLimit-Remaining: 50 RateLimit-Reset: 60¶
The key runtime value is the first element of the list: expiring-limit
, the others quota-policy
are informative.
So for the following header:¶
RateLimit-Limit: 100, 100;w=60;burst=1000;comment="sliding window", 5000;w=3600;burst=0;comment="fixed window"¶
the key value is the one referencing the lowest limit: 100
¶
The most common syntax we found on the web is X-RateLimit-*
and
when starting this I-D we opted for it¶
The basic form of those headers is easily parseable, even by implementors procesing responses using technologies like dynamic interpreter with limited syntax.¶
Using a single header complicates parsing and takes a significantly different approach from the existing ones: this can limit adoption.¶