- Newest
- Most votes
- Most comments
No downtime unless you're making a destructive change (i.e. deleting resolver or data source").
It should be noted that while update operations do not have downtime, they are also not instantaneous. There could be a short amount of time (a few minutes) after performing an update where the function or resolver is still using the old code. Consider the following scenario:
- AppSync resolver v1 points to lambda function v1
- A new release is created and deployed and thus the AppSync Resolver is updated to V2 and the lambda function is simultaneously updated to v2 (both update commands return ok)
- Request for the AppSync resolver are being handled by V1 for the next ~1 minute. If lambda function V2 is not compatible with resolver V1, there could be issues
- AppSync infrastructure finishes propagating the V2 changes and all requests are handled by the V2 resolver
So although the AppSync update API call looks like it returns completed in step 2, there are asynchronous activities by the AWS backend which may cause the changes to AppSync to not become active for some time. The time between 3 and 4 above could be a number of minutes during which any requests to the AppSync API from consumers could have unpredictable responses.
answered 2 years ago
Relevant content
- AWS OFFICIALUpdated 2 years ago

I have a problem that there is a downtime between removing a resolver then replacing it with another resolver, is there a solution for the downtime here?