Mathieu Poussin

What would be the easiest way to extract the list items ? I only get the very long line from the logs but it doesn't looks like it's easily parseable...

I can reproduce it in our CI, I will check if I can also reproduce it locally (it did not happen during my local tests)

Sorry for the delay, This is what I get from the logs (Also I had to add the missing timeout parameter to `force_flush` method) ```json {"level":"WARN","timestamp":"2023-01-18T13:06:26.900+00:00","logger.method_name":"cardiologs.open_telemetry","message":"Array, OpenTelemetry::SDK::Trace::Span"} ``` Somehow an...

I got some progress adding some recursive function ```ruby def scan_array(given_array, level: 1, max_level: 20) throw Exception.new("Too many levels") if level > max_level batch_classes = given_array.map { |i| i.class }.uniq...

More investigation focusing on the rake task. I do see a strange behaviour, the same Rake task `fake_data` is showing 2 different spans, one in the batch and one in...

[spans.log](https://github.com/open-telemetry/opentelemetry-ruby/files/10448397/spans.log) I have attached the span logs from my code. To achieve this, we are running multiple rake jobs in a single rake call: `bundle exec rake db:create db:schema:load setup_db:fake_data`...

Ok I got a reproducible scenario (not isolated from my codebase but it should be easy to isolate) Make sure you you have the OTLP exporter configured (but no target...

Let me know if you have issues reproducing it

I checked I could find both spans related to PostgreSQL, ActiveRecord and Redis having this similar issue (which are most of the spans our apps generate, so it's possible that...

It's not the first time I see strange behaviour under high load, maybe it's related: https://github.com/open-telemetry/opentelemetry-ruby/issues/1418