Pass correlation information to Elasticsearch bulk requests to correlate problematic updates
It is currently difficult to isolate tasks that that can be problematic during a bulk index.
Example: 2020-06-23: High CPU on Elasticsearch Advanced Global Search nodes
In order to make this easier to Identify in the future We should be passing our labkit correlation IDs through as an X-Opaque-Id
header.
Then we can later track a long running bulk update back to the worker and hence project in GitLab. This could considerably narrow the search space for problems.
Edited by Dylan Griffith