Issue/jperf 566 fix ci flakiness

[dagguh]

No description provided.

[dagguh]

Here's 4 runs of da9abac :

• https://circleci.com/gh/dagguh/docker-infrastructure/7
• https://circleci.com/gh/dagguh/docker-infrastructure/8
• https://circleci.com/gh/dagguh/docker-infrastructure/9
• https://circleci.com/gh/dagguh/docker-infrastructure/10

It still flakes. I could bump the timeout even further, but there is something fishy here. Looking in the stdout of the tests from the test reports in the artifacts we can look for the org.apache.catalina.startup.Catalina.start Server startup in line. We see two such lines, due to two test methods. One reports in 11-13 seconds and the other (docker-to-docker) is 170-280 seconds. And then yet again, one times out completely despite an 8 minute timeout.

I'll try to replace the log polling with GUI polling

[dagguh]

https://circleci.com/workflow-run/fa542720-2a09-40a1-ba5e-fc00fc432c4f
This flake is the second kind: it's unhappy with the admin fullname field.
This blocks me from merging atlassian/jira-actions#29 and I have no idea how to fix it.

[dagguh]

This is still not working, I'll close and reopen when CircleCI goes green

[dagguh]

The video from the CI showed that the license setup takes around 60 seconds and our timeout for the next screen is also 60 seconds. I bumped the WD timeout and now it's all green. Even after a full 4x rerun or two.
The extra timeout means the CI runs for a little longer.
So this trades a minute (or two) of dev loop for reliability. In total, this should save a lot of time, due to no need for reruns.

[wyrzyk]

😱

[dagguh]

That's the ExpectedCondition SPI