Log to Metrics filter use container OOM Kill about fluent-bit HOT 5 CLOSED

Hi there. Are you consuming the data from the exporter? Does it reproduce is you use another output? (stdout, null, http...)

from fluent-bit.

Maybe related to this: #8422

from fluent-bit.

Maybe related to this: #8422

Thanks for letting us know.
I'll try it and let you know when it's available in 3.0.0~!

I'm testing it by adjusting various options in Fluent-bit, but still causing OOM whenever there's a lot of logs generated at once due to instant traffic...

from fluent-bit.

Hi there. Are you consuming the data from the exporter? Does it reproduce is you use another output? (stdout, null, http...)

Hi, @lecaros
We're currently exposing through the prometheus_exporter, which is shown as an example in the official documentation, and registering as a servicemonitor in prometheus to get the metrics.

I haven't tried other outputs, but when the traffic is low, the metrics are generated normally, but when the traffic is high and the logs are large, Fluent-bit causes a memory leak.
(I'll test with other outputs and let you know.)
Looking at this, the output, prometheus_exporter, seems to be working fine, but the GC is not working properly.

from fluent-bit.

Hi, @drbugfinder-work
I bumped it up to 3.0.0 and it seems to have fixed the Fluent-bit OOM issue. I will add the plugin you mentioned and test it and share it with you.

from fluent-bit.