Open ralphbean opened 3 days ago
The current retry strategy is to retry three times with exponential backoff at 1sec, 2sec and 4sec with a bit of random jitter.
The overall execution of the CLI has a default timeout of 5min, that can be configured using the --timeout
parameter.
@ralphbean do you have a suggestion as to how we can tweak this defaults?
No, no suggestion yet. That strategy looks reasonable.
Describe the Bug
When trying to validate an artifact at release time, ec-cli failed with violations that indicate that it timed out trying to talk to quay. This happened 3 or 4 times to me in the last week.
I suspect quay configuration changes may be causing the increase in frequency, but I cannot prove it.
Steps to Reproduce
I cannot reproduce. :(
Expected Behavior
Successful ec validation.
Actual Behavior
ec reports:
Screenshots or Terminal Output
N/A.
Environment Details
This happened in the konflux release-serivce pipelines.
Possible Solution
If quay is down, you can't help it. But, you can retry! Can you evaluate your retry mechanism to improve resilience of the end to end process?