test: increase timeouts for tests that are flaky on slow CI

[Frando]

Description

We have some tests that use timeouts to not wait infinitely for an event that might not be coming. These tests are flaky if the timeout is too low, especially on windows and likely if the machines are overworked. This PR increases these timeouts:

• Increase timeout of test_node_add_tagged_blob_event from 1s to 10s (Fixes Flakey: node::tests::test_node_add_taggded_blob_event #2331)
• Increase timeouts of the pkarr_publish_dns_resolve_* tests from 2s to 10s (Fixes flaky test: iroh-net discovery::test_dns_pkarr::pkarr_publish_dns_discover  #2221)

Breaking Changes

Notes & open questions

Change checklist

• Self-review.

[dignifiedquire]

maybe log the time it took, so we can see a little better if this was really the issue and what variance is in these

[Frando]

maybe log the time it took, so we can see a little better if this was really the issue and what variance is in these

Sure, we can add that, but will we ever read the logs of successful CI runs for these? Because logging the time doesn't help on the failed runs, right? We'd need something like "CI metrics collection" to actually make use of these numbers. For general variance, I'd say total test runtime is already good enough likely.

[flub]

maybe log the time it took, so we can see a little better if this was really the issue and what variance is in these

Sure, we can add that, but will we ever read the logs of successful CI runs for these? Because logging the time doesn't help on the failed runs, right? We'd need something like "CI metrics collection" to actually make use of these numbers. For general variance, I'd say total test runtime is already good enough likely.

Yeah, I agree this is not terribly effective. I don't care about it really.

[dignifiedquire]

Because logging the time doesn't help on the failed runs, right?

well I would look at it retrospectively, if this is close to the timeout often, or not