Improve resilience against ES log collection issues in CI contexts

[soxofaan]

From time to time openeo-geopyspark-driver unit tests or integration tests fail because of ES log collection failures. For example something like

Job error logs: [{'id': '', 'code': 'Internal', 'level': 'error', 'message': "Log collection for job j-231102dbf0e4490692a4a7dc66ada773 failed. (req_id: r-23110211582f43ac92e799380c550328) OpenEOApiException(status_code=504, code='Internal', message='Temporary failure while retrieving logs: ConnectionTimeout. Please try again and report this error if it persists. (ref: no-request)', id='no-request')"}]

In most cases, the error logs are not relevant in any way for the test itself, and failure to get them due to uncontrollable root cause should not fail the whole test.

[soxofaan]

e.g. @JeroenVerstraelen you added Open-EO/openeo-geopyspark-integrationtests@abed154 in the integration tests which automatically fetches error logs (the job.logs(level='ERROR') there) when something is unexpected, but when getting the error logs fails the original issue is obfuscated.

[soxofaan]

related:

• revert assert_batch_job helper openeo-geopyspark-integrationtests#12