L15. Monitoring and Observability: Cloud Monitoring, Logging, Trace, and Error Reporting

Google Cloud Operations Suite

Formerly known as Stackdriver, Google Cloud Operations Suite is the integrated observability platform for Google Cloud and hybrid environments. Four pillars of observability:

1. Metrics (what happened numerically)
2. Logs (what events occurred with context)
3. Traces (how requests flowed through services)
4. Errors (what went wrong in code)

Cloud Monitoring

Cloud Monitoring collects metrics from Google Cloud services, custom applications, and third-party services. Key features:

• Pre-built dashboards for all Google Cloud services
• Custom dashboards and metric explorer
• Alerting policies: trigger notifications (email, PagerDuty, Slack, etc.) when metrics breach thresholds
• Uptime checks: synthetic monitoring to verify service availability from multiple locations worldwide
• SLO monitoring: define and track Service Level Objectives

Metrics are free for Google Cloud services; custom metrics have a charge above a free tier.

Cloud Logging

Cloud Logging collects, stores, and analyzes log data from Google Cloud services and custom applications. Log types:

• Platform logs: automatically generated by Google Cloud services (Compute Engine, GKE, Cloud SQL)
• User-defined logs: custom logs written by your applications
• Audit logs: who did what, when, from where (Admin Activity, Data Access, System Events)

Log Router: route logs to Cloud Storage (long-term archival), BigQuery (analysis), or Pub/Sub (streaming). Log Explorer: search and filter logs in real time using structured queries.

Cloud Trace

Cloud Trace is a distributed tracing system that shows how requests propagate through your application and its dependencies, including latency breakdown. Use for: identifying performance bottlenecks, understanding request flow across microservices.

Error Reporting

Error Reporting aggregates application error messages, groups them, and notifies when error rates spike. Supported languages: Go, Java, Python, Node.js, PHP, Ruby, .NET, C++.

Cloud Profiler

Cloud Profiler continuously gathers CPU and memory usage profiles from production applications to identify performance hotspots without impacting performance.

Exam tip: Cloud Monitoring = metrics and alerting. Cloud Logging = structured log analysis and Audit Logs. Cloud Trace = latency and request flow tracing for microservices.