A collector cluster shows `otelcol_exporter_queue_size` consistently near `otelcol_exporter_queue_capacity`, and `otelcol_exporter_send_failed_spans` is also increasing because the backend is overloaded. What is the best interpretation?

Question

Accepted Answer

B. Scaling the collector alone is unlikely to help because the backend is the bottleneck, and adding pressure may worsen the situation. The scaling guide says queue pressure can signal exporter slowdown, but it also warns that persistent send failures and backend bottlenecks are cases where scaling collectors alone will not help and may worsen pressure on the backend. That is the documented `when NOT to scale` scenario. Option a is wrong because queue growth is not always a collector-capacity problem; it can reflect a slow or failing destination.

A collector cluster shows `otelcol_exporter_queue_size` consistently near `otelcol_exporter_queue_capacity`, and `otelcol_exporter_send_failed_spans` is also increasing because the backend is overloaded. What is the best interpretation?

Related Questions

Discussion

A collector cluster shows otelcol_exporter_queue_size consistently near otelcol_exporter_queue_capacity, and otelcol_exporter_send_failed_spans is also increasing because the backend is overloaded. What is the best interpretation?

Related Questions

Discussion

A collector cluster shows `otelcol_exporter_queue_size` consistently near `otelcol_exporter_queue_capacity`, and `otelcol_exporter_send_failed_spans` is also increasing because the backend is overloaded. What is the best interpretation?