Old behavior
When an app gives an error response (status code 4xx or 5xx) while refreshing the OAuth2 access token, Zapier keeps retrying the Zap step indefinitely or until it hits a certain limit, depending on the user’s settings.New behavior
When an app encounter a 4xx error response (except for the ones listed below) while refreshing the access token, Zapier will mark the connect as stale, and send an email telling the user to reconnect. Exceptions: The following 4xx errors often indicate a temporary issue so they still have the same behavior as before:- 408 (Request Timeout)
- 409 (Conflict)
- 423 (Locked)
- 425 (Too Early)
- 429 (Too Many Requests)
- If the stale connection is used by a trigger step, the trigger polling system will skip polling when the scheduled time comes.
- If the stale connection is used by an action step, the Zap run will be put on hold until the user reconnects and replays the run.
What does it mean to you?
You don’t need to handle 4xx error responses inrefreshAccessToken
anymore. For example, you might have been catching 4xx errors in refreshAccessToken
by enabling skipThrowForStatus
and throwing ExpiredAuthError
: