Hello, We are deployed the signoz

otel-collector-metrics

in our environment, we have noticed that the pod gradually increases in memory usage and once it reaches the resource limit it is OOMKilled. We have tried multiple options with the batch size: • 50k batch size, 30s timeout • 20k batch size, 15s timeout • 50k batch size, 10s timeout But nothing seems to be working, one way was to reduce the number of scrape job configurations in the prometheus receiver but that was temporary and it would still increase in memory utilisation.

It is not scalable to scrape using

otel-collector-metrics

for such cases. Please look into https://opentelemetry.io/docs/kubernetes/operator/target-allocator/

Got it, thank you!