Tests are flakey?

[sublimator]

Are the integration tests flakey?

Suggest:

• Increase Github Actions test runner size (more compute/memory)
• Add a retry mechanism - simply retrying tests a few times, while not ideal, makes the tests much more useful in practice

@mvadari ?

[justinr1234]

@sublimator would you be able to link a couple logs or show more specific info?

[sublimator]

It's a question. I thought I noticed a PR build failed different jobs at different times. On mobile, so I don't have logs. If they are known flakey(as integration tests can be), I will pay them less heed.

[justinr1234]

Got it, I’m not aware of them being flakey at the moment. The last time they were was when a faucet was going out of commission, but we fixed that.

If you notice any flakey, please drop a line on here.

[sublimator]

Will do, thx the info!

[sublimator]

Here was where I started wondering:
#2273 (comment)

[sublimator]

This looks related to the faucet you mentioned:

However, in a previous job the "browser" job failed rather than one or more of the "integration" jobs

[mvadari]

It's known that the tests are slightly flaky, but they appear to be failing more than usual on your PR. It's almost always because of faucet flakiness/rate limiting.

[sublimator]

5th time's a charm!

Can see a few other PRs that seem at a glance to be victims:

• doc: Make the docs more robust with examples #2226
• add submitAndWaitBatch sugar function #2149

[sublimator]

Must have had bad luck before.
Pushes since opening this haven't yielded any issues.
Closing.