Carson Ip

Still needs a bit of work to make the benchmarks work because benchmark measure generated count instead of exported doc count. Since sending_queue is enabled, the rate of Consume* will...

> @carsonip should we disable the queue for benchmarks, so they're focused just on the ES exporter perf? I believe it will still break the benchmarks because batch sender will...

Benchmark result compared to main with #33621 : ``` goos: linux goarch: amd64 pkg: github.com/open-telemetry/opentelemetry-collector-contrib/exporter/elasticsearchexporter/integrationtest cpu: 11th Gen Intel(R) Core(TM) i7-11800H @ 2.30GHz │ main.out │ batchsender.out │ │ sec/op...

> Also an important point to note: with this PR the docs/s is not always equal to events/s, which may indicate a bug. Will dig into it later. There was...

> Benchmark result compared to main with https://github.com/open-telemetry/opentelemetry-collector-contrib/pull/33621 : I looked into the perf regression and figured out for root cause of it. It is because on `main`, events are...

> playing devil's advocate: do we want to fully use CPUs in the exporter side, or is there an advantage to tying up only one goroutine on encode? Are there...

Superseded by #34238

Closing as completed via https://github.com/open-telemetry/opentelemetry-collector-contrib/pull/34238 . Requires opt-in for the new behavior.

hi @tmaiaroto, I am using this theme in my own blog and therefore I can take up some of the work about this theme. I can even maintain the repo...

@pboling Although I am still using this theme, I am not a maintainer of the theme. It has been a long time since the last time I touch any of...