Include wait time in more cases when UI testing progress bubble color

[Erin007]

Related to X-TEAM 366

While investigating UI test failures on a Drone run I stumbled on what appears to be some flakiness in the tests covering multi page assessments. In the Drone run and then occasionally locally on a fresh pull of staging there was an error where we expected a bubble to be purple (perfect), but it was showing as white (not tried).

There was a time-saving optimization in place to only add additional wait time if the status of the bubble was not_tried, but I think the additional wait time is needed in other cases as well to reduce flakiness in this test. These tests consistently pass for me locally when the additional wait time and completion of Ajax requests are added in as steps.

[Erin007]

I don't think any of the LP tests related to bubble color have hit our flakiness threshold, but it does look like there have been re-runs needed on a few with a similar error in the past couple of days.

[jamescodeorg]

Thanks for digging into this! I'm fine with switching to this algorithm, but I'm not sure I understand why this change helps.

The original code was supposed to wait up to 30s for the bubble to turn colors. From the logs, it doesn't look like this polling happened (i.e. the test gave up after the first immediate failure)?

[Erin007]

Thanks for digging into this! I'm fine with switching to this algorithm, but I'm not sure I understand why this change helps.

The original code was supposed to wait up to 30s for the bubble to turn colors. From the logs, it doesn't look like this polling happened (i.e. the test gave up after the first immediate failure)?

Hmmm. I'm not 100% sure either. There are a few other tests that use the wait_short_until step. I'll take a peek to see if it's working in those scenarios.