-
Type:
Task
-
Resolution: Unresolved
-
Priority:
Unknown
-
None
-
Component/s: AI/ML
-
None
-
Not Needed
We have seen frequent failures when testing against the remote Atlas cluster, including timeouts due to contention between tests running on the same cluster (INTPYTHON-760), and host issues like certificate verification failure (DRIVERS-3306).
We had faced similar issues of instability in Drivers Atlas Testing, and had used workarounds like staggering the tasks and adding retries.
Instead, I propose that we only test against the Atlas Local docker image. Currently the one known behavioral difference is that the search index update command does not work (DRIVERS-3078). We could consider blocking this ticket on that one.
- is related to
-
DRIVERS-3306 [AI/ML Pipeline] Always attempt to connect to the cluster
-
- Backlog
-
-
INTPYTHON-760 [ai-ml-pipeline-testing] Add docs and policy around separate clusters
-
- Ready for Work
-
-
DRIVERS-3078 Include the type field in updateSearchIndex command document
-
- Backlog
-