- feat: add logLevel and logRetryAfterSeconds options when initializing the lib (#242) 8da6526
- fix: APAC region documentation and test expectations (#240) fbcf3c2
- fix: update unauthorized e2e test for latest breaking change (#228) 56ec324
- fix: list replication lag test failures (#225) d974951
- chore: update all dependencies (#215) 4cf1346
depends on:
logRetryAfterSeconds option
🔔 aio-lib-state and aio-lib-core-networking feature release: if a network call is throttled by a 429 http status response, it will also send a Retry-After header. aio-lib-state (which uses aio-lib-core-networking) will retry the call according to this header automatically (with exponential backoff). However, there will be cases where the cumulative results of the retries will exceed the timeout limit of the action itself (which is 60s) but there is no indication that the cause of the timeout is because of the retries.
This feature release adds a logRetryAfterSeconds option in both libraries, and it will default to 10 seconds. If any Retry-After header value (from the result of the networking call) exceeds this option value, it will always log the retry as a warning to surface this issue where you might be exceeding your usage for the State service but you did not know you were being throttled.
Previously it will only log the retry only if the logging level was set to debug. To disable this behavior, you can set logRetryAfterSeconds to 0 explicitly.