Because the Google Chat API is a shared service, we apply quotas and limitations to make sure that it's used fairly by all users and to protect the overall performance of Google Workspace.
If you exceed a quota, you'll receive a 429: Too many requests HTTP
status code response. Additional rate limit checks on the Chat
backend might also generate the same error response. If this error happens, you
should use an
exponential backoff algorithm
and try again later. As long as you stay within the per-minute quotas listed in
the following tables, there's no limit to the number of requests you can make
per day.
Two quota types apply to Chat API methods: per-space and per-project quotas.
Per-space quotas
Per-space quotas limit the rate of queries in a given space and are shared among all Chat apps acting in that space calling the listed Chat API methods for each quota.
The following table details per-space query limits:
Per-space Quota |
Chat API methods |
Limit (per 60 seconds, shared |
|---|---|---|
Reads per minute |
|
900 |
Writes per minute |
|
60 |
Per-project quotas
Per-project quotas limit the rate of queries for a Google Cloud project, and thus apply to a single Chat app calling the specified Chat API methods for each quota.
The following table details per-project query limits. You can also find these limits on the Quotas page.
Per-project Quota |
Chat API methods |
Limit (per 60 seconds) |
|---|---|---|
Message writes per minute |
|
3000 |
Message reads per minute |
|
3000 |
Membership writes per minute |
|
300 |
Membership reads per minute |
|
3000 |
Space writes per minute |
|
60 |
Space reads per minute |
|
3000 |
Attachment writes per minute |
|
600 |
Attachment reads per minute |
|
3000 |
Reaction writes per minute |
|
600 |
Reaction reads per minute |
|
3000 |
Additional usage limits
There are additional quota limits for creating spaces of type GROUP_CHAT
or SPACE (by using either the spaces.create or the spaces.setup method).
Create less than 35 spaces per minute and 210 spaces per
hour of these types. Spaces of type DIRECT_MESSAGE aren't subject to these
additional quota limits.
Large queries per second (QPS) of any API targeting the same space can trigger additional internal limits that aren't visible in the Quotas page.
Resolve time-based quota errors
For all time-based errors (maximum of N requests per X minutes), we recommend your code catches the exception and uses a truncated exponential backoff to make sure your devices don't generate excessive load.
Exponential backoff is a standard error handling strategy for network applications. An exponential backoff algorithm retries requests using exponentially increasing wait times between requests, up to a maximum backoff time. If requests are still unsuccessful, it's important that the delays between requests increase over time until the request is successful.
Example algorithm
An exponential backoff algorithm retries requests exponentially, increasing the wait time between retries up to a maximum backoff time. For example:
- Make a request to Google Chat API.
- If the request fails, wait 1 +
random_number_millisecondsand retry the request. - If the request fails, wait 2 +
random_number_millisecondsand retry the request. - If the request fails, wait 4 +
random_number_millisecondsand retry the request. - And so on, up to a
maximum_backofftime. - Continue waiting and retrying up to some maximum number of retries, but don't increase the wait period between retries.
where:
- The wait time is
min(((2^n)+random_number_milliseconds), maximum_backoff), withnincremented by 1 for each iteration (request). random_number_millisecondsis a random number of milliseconds less than or equal to 1,000. This helps to avoid cases in which many clients are synchronized by some situation and all retry at once, sending requests in synchronized waves. The value ofrandom_number_millisecondsis recalculated after each retry request.maximum_backoffis typically 32 or 64 seconds. The appropriate value depends on the use case.
The client can continue retrying after it has reached the maximum_backoff time.
Retries after this point don't need to continue increasing backoff time. For
example, if a client uses a maximum_backoff time of 64 seconds, then after reaching
this value, the client can retry every 64 seconds. At some point,
clients should be prevented from retrying indefinitely.
The wait time between retries and the number of retries depend on your use case and network conditions.
Request a per-project quota increase
Depending on your project's resource usage, you might want to request a quota increase. API calls by a service account are considered to be using a single account. Applying for an increased quota doesn't guarantee approval. Large quota increases can take longer to be approved.
Not all projects have the same quotas. As you increasingly use Google Cloud over time, your quotas might need to increase. If you expect a notable upcoming increase in usage, you can proactively request quota adjustments from the Quotas page in the Google Cloud console.
To learn more, see the following resources: