Open harrdou opened 1 year ago
This might be an issue with the new RPConfig cache, or it might be an issue with our session cookie being blocked.
I've disabled the RPConfig cache in CATE and will monitor.
We should also add some error checking in strategic spots to log cases when a session is mysteriously not found.
Definitely seems to be an issue with the RPConfig cache. Disabled it for now. Will see if the issue goes away after the move to AWS ElastiCache for Redis.
I wouldn't close this ticket if we want to monitor it, thought?
I wouldn't close this ticket if we want to monitor it, thought?
Azure Boards had more states for bugs. New - > Investigate > Approved -> Fixed -> Acceptance Test Pass.
Not sure how this works with regular GitHub.
Ill re-open it for now, and will figure out how the board mapping works
The issue hasn't re-occurred in CATE since we turned the cache off there on the 14th.
We could revert to the old implementation but that one made RP config changes a pain because he cache had to be flushed manually by forcing a re-load of the interception scripts. The performance gain offered by the cache is probably not worth it today anyway, given our current workload. Once we get to load testing the new infrastructure we can examine whether it makes sense to re-introduce it.
sounds good, how do we ensure the cache will be off in production?
Hand-offs to the TOTP service are failing in prepareForStep because the mfaProvider is not being found in the RPConfig...
weakmfa: prepareForStep. mfaProvider is missing!