Observability

This reference covers observability on Suga: viewing logs, monitoring metrics, and debugging issues.

Logs

Logs have their own reference page. See Logs for structured logging, severity detection, library compatibility, retention, and limitations.

Access metrics from the Metrics tab in the service properties panel.

Shows CPU cores consumed over time.

Pattern	Interpretation
Consistently > 80%	Increase CPU allocation
Consistently < 20%	Consider decreasing (save costs)
At 100%	Throttled, users experiencing slowness

Shows MiB/GiB consumed over time.

Pattern	Interpretation
Consistently > 80%	Risk of OOMKilled, increase memory
Growing steadily	Possible memory leak
Consistently < 30%	Over provisioned

Memory at 90%+ means high risk of OOMKilled crashes. Increase memory allocation.