Seth Gupton
Seth Gupton
Thank you @mfbieber ! Looking forwards to this getting merged!
Redis should quick... But I'm also wondering why does IsTokenRevoked need locks around it at all? If there is already a latency of 1 hour around between calls to loadRevokedTokens...
I found it! A '-1' is being passed in for the Scan count. https://github.com/argoproj/argo-cd/blob/master/util/session/state.go#L84 '-1' causes it to use the default of 10. https://github.com/redis/go-redis/blob/master/generic_commands.go#L362-L364 We have over 4 million keys...
Well wasn't instantaneous but it did bring it from 5 - 15 minutes down to 11 to 17 seconds.
Decreased these hourly spikes in redis requests;  But increased spikes in redis duration;  I am unsure if that is anything to be concerned about. Right now we're happy...
Just stumbled face first into this too. Hooks running through the postrenderer would be great.
Notes from contributors meeting: Try to see if there is a number between 10 and 10000 that works nicely. Prefer not adding a new configuration value to avoid having too...
I also ran into this. Made it really confusing when there are multiple helm charts in a Warehouse.
P.S. We are solving this in a different way right now. But it would be nice to have it built in. ``` - name: Generate random retry_wait_seconds id: randomize-retry run:...