Make e2e tests more thorough

[emmatyping]

This tests code deeper in the pipeline that might have caught #1668

Task list:

1. Add test for symbolification pipeline, based on test in Sentry repo #1888
2. Add test for alerts/notifications to test the snuba pipeline #1887

[BYK]

@ethanhs the backup/restore dance is woefully undertested. I did some fixes and improvements there while writing the docs (it did not even work before that 😅). In the upgrade scenario, you may wanna add a backup/restore scenario too. That said it should only matter when there's a model/migration change.

[emmatyping]

Good idea! I think we've seen a couple of tickets about backup/restore, so it would be good to test that.

[emmatyping]

Arpad suggested for testing symbolicator workflows:

That seems like a good idea, yes. To cover this particular usecase, you would need a test that uses all the builtin symbol servers. A Windows minidump perhaps. We should have one already in the main Sentry tests, though those tests are always testing the very latest symbolicator master.

[chadwhitacre]

backup/restore

Punting to getsentry/sentry#36868, need to address more comprehensively.

[chadwhitacre]

What else could/should we test?

SSO & integrations ... would involve many more moving parts, more difficult to test.

[chadwhitacre]

Profiling should come in with #1838.

Replays (#1873)? Dynamic Sampling?

[chadwhitacre]

FE? Not really worth doing here, leave that in sentry.

We want to exercise the containers, that's really the scope.

Do we have tests for updating settings?

[chadwhitacre]

Closing with 1 of 2 completed.