Observe Tool Use Cloud Logging + Logs Explorer central logs; write JSON to stdout for structured fields (severity, trace) Log-based metrics count matching log entries (e.g. payment failures) Cloud Monitoring metrics, dashboards, uptime checks, alerting policies, SLOs Cloud Trace / Profiler latency traces (OpenTelemetry) and continuous profiling Error Reporting groups exceptions automatically Log sinks route logs to BigQuery / Storage / Pub/Sub (and exclude noisy logs) Control cost Budgets & alerts per billing account/project, optionally publishing to Pub/Sub for automation. Billing export to BigQuery for detailed analysis by label/service/SKU. Labels on every resource (env, team, app). Recommender for idle VMs, over-provisioned resources and unused IPs; committed use discounts for steady…