@vsizov Instrumentation indeed requires the whole Rails application, not just for the source code but also for the InfluxDB settings. Can we perhaps turn the indexer into a periodically executing Sidekiq worker? In that case the worker will be instrumented automatically and there's nothing extra we'd need to do.
@yorickpeterse Actually it's already a Sidekiq job I mean Sidekiq job runs the indexer as a separate process and wait the results. We have done it in this way because of lots of memory leaks in indexer. BUT if we wait the indexer we can probably just use Sidekiq instrumentation just like you said because we don't need too many details like how much does it take to index commits or blobs. So thanks for the idea.
Thanks @vsizov I think that we need to break the blackbox that elasticsearch execution is in way that we can track what's it doing at a given time - what's being indexed, how much time it takes, did it failed? if so, why? etc.