tofu-controller
tofu-controller copied to clipboard
Incorrect gotk_reconcile_condition metric after migration to v0.16.0-rc.3
Hi,
I've upgraded from v0.14.0 to v0.16.0-rc.3.
The gotk_reconcile_condition for the Ready condition is not being correctly reported.
- Terraforms are Ready
- Metric reports Unknown=1 and True/False=0
Do you have any idea the cause? Could you please help me troubleshoot this issue?
Thank you in advance.
More details:
v0.16.0-rc.3
v0.14.0
Ran the controller with log-level=trace:
Thank you @renato0307 for reporting this.
Maybe it's because we have recently changed the code base to match that of Flux v2 GA.
@luizbafilho @yitsushi could you help take a look please?
cc @LappleApple
Anyone working on this @LappleApple @chanwit? May I give it a go?
Hey @renato0307, all yours. :) I'll assign it to you.
@renato0307 : Go for it, if any of us picked it up, we assign ourselves.
Looks like the problem is on the defer.
The defer uses the original terraform var which contains the Unknown condition.
It is overwriting the metric written here, using the reconciledTerraform var.
Hi team, please evaluate this fix.
Thank you for bringing this up @adonispd could anyone please help use clarify this bug again? Happy to help review and land a fix if someone could point things out clearly.
Hi @chanwit, I could reproduce this issue in the v0.16.0-rc.4. Wrote a short summary on the PR. #1287
@chanwit any idea when the team will create a new release candidate?