You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the order of operations is annotation > watches > command-line flag. I'd recommend we make the command line flag able to override what is set in watches. Watches is baked into the container and somewhat immutable in the runtime environment, whereas the deployer can set command line flags for per-deployment behavior. Similarly, the CRD annotation is useful per object.
In the documentation, we advise against mixing these. I'm not sure why we do this. Can we remove that text or provide more explanation why that's a bad idea to mix the two.
In the log, there's no indication when the next reconciliation can be expected (worse case) or the effective value. Should this value be part of the message or KV pairs that are logged?
The text was updated successfully, but these errors were encountered:
Nevermind on this, I can see how there can be multiple watches and using the command-line flag as the override could bigger impact than when operator is 1 for 1.
Having different issue where reconcilePeriod is getting picked up (10h), but still reconciling every minute on 1.32 and 1.33 and trying to run that down.
reconcilePeriod
Want to propose:
Currently, the order of operations is annotation > watches > command-line flag. I'd recommend we make the command line flag able to override what is set in watches. Watches is baked into the container and somewhat immutable in the runtime environment, whereas the deployer can set command line flags for per-deployment behavior. Similarly, the CRD annotation is useful per object.
In the documentation, we advise against mixing these. I'm not sure why we do this. Can we remove that text or provide more explanation why that's a bad idea to mix the two.
In the log, there's no indication when the next reconciliation can be expected (worse case) or the effective value. Should this value be part of the message or KV pairs that are logged?
The text was updated successfully, but these errors were encountered: