Yuri Shkuro

> as Jaeger doesn't provide functionality to search in log fields It certainly does with Cassandra, and I thought with ES too

@paule96 the dependencies calculation done by all-in-one happens inside `memory` storage: https://github.com/jaegertracing/jaeger/blob/711488f94f25d4a5e7900d0ba450ecb5bfdfda7b/plugin/storage/memory/memory.go#L63 It may not be very straightforward to combine this with the enricher idea because dependencies logic does not...

@dependabot ignore this minor version

I suggest you look up some tuning guides for ES.

What is your evidence that this aspect of the queue has anything to do with CPU usage or dropped spans?

Sorry, I don't really follow your experiment. Are you saying that you keep the queue capacity fixed (at 1M items), but vary the batch size of the items in the...

I am not sure I follow your explanation. If you use the same _total_ queue capacity (i.e in case of batched queue you flatten the batch for the purpose of...

You're not using memory-only version, you're running with Badger which writes to files. Something may be getting corrupted when it OOMs. You need to configure data purging for Badger to...

I am not an expert on Badger, but there is a TTL flag we expose: https://www.jaegertracing.io/docs/1.28/cli/#jaeger-all-in-one-badger

> Support repeated tags/logs keys. why not unique tags instead?