correlation_id is not maintained across Sidekiq retries
At present, every worker retry will log messages with a different correlation id. This makes it cumbersome to search for related log entries. In order to get the original correlation id you currently have to:
- Get the JID of the retry, since the JID is constant across retries
- Filter for that JID, go to the first time the job was enqueued
- Get the correlation_id of that first job
We should consider instead keeping the correlation_id constant across retries