test_compatibility snapshot is flaky

[koivunej]

https://neon-github-public-dev.s3.amazonaws.com/reports/pr-7639/9034227206/index.html#suites/0fbaeb107ef328d03993d44a1fb15690/463911a4a21e1ef3

Because of timeout waiting for the Lsn in create snapshot. I don't understand how can this happen, we were stuck 10s at the same Lsn, pageserver went through all of the safekeepers. After the failure sync safekeepers the safe safekeepers were above the waited Lsn.

Marking this for pageserver so we might triage this, but this might be something interesting to look for compute team. Or maybe I missed something obvious.