rate-limits.md

Rate limits

Rate limits are defined by each Mastodon server, and may vary. Sensible defaults are defined here: https://docs.joinmastodon.org/api/rate-limits/

Consensus Chess nodes attempt to be good citizens by rate limiting themselves. To this end, they track usage of the APIs with a RateLimiter object, and self-limit (adding delays to communications) if they exceed their allotted usage.

Usage is defined as:

• int permitted - number of permitted requests in the time period
• TimeSpan period - period within which number of requests is measured

Mastonet does not (yet) expose rate limiting headers from the Mastodon API. I've filed issue #85 - expose rate-limiting headers to ask about it.

Monitoring

If an instance is rate-limiting itself, it'll log a warning message to that effect, and delay the delivery of its message:

Rate limit exceeded... Introducing a delay.

Rate limiting is a synchronous activity. If one message is delayed, all other messages are delayed too.

Metrics

At current time, there's no queue or management for delayed messages, nor is there a metric to help track the number of currently delayed messages.

See:

• ICG-75 - Implement message and API activity queues to better manage and observe rate limits