Google Cloud Monitoring

Overview

Google Cloud Monitoring (formerly Stackdriver) is Google Cloud's native observability platform, launched in 2010 under the Stackdriver brand and eventually unified into Google Cloud's monitoring solution. Cloud Monitoring provides automatic metric collection across all GCP services, alerting policies, dashboards, Uptime Checks, and Service Monitoring. Cloud Logging offers centralized log management with real-time streaming analysis. Error Reporting automatically aggregates and analyzes application errors. Cloud Monitoring integrates deeply with Vertex AI for AI/ML model monitoring and MLOps scenarios. It is the standard monitoring solution for GCP-centric cloud-native teams.

Key Strengths

• All GCP Services Pre-Integrated: Compute Engine, GKE, Cloud Run, BigQuery, and every other GCP service automatically publishes metrics to Cloud Monitoring — monitoring starts when you create resources. See Website Monitoring Tools Guide.
• Logs-Metrics-Errors Unified: Cloud Monitoring, Cloud Logging, and Error Reporting are deeply integrated, enabling cross-correlation across logs, metrics, and error data without tool switching.
• Managed Prometheus Natively: GCP offers Managed Prometheus — a fully managed Prometheus service compatible with PromQL and Grafana dashboards. Ideal for teams already using the Prometheus standard. Combine with monitoring best practices.
• Flexible Alerting Policies: Cloud Monitoring alerting supports threshold-based, metric-absence, PromQL-based, and log-based conditions in flexible combinations, reducing false alarms.
• Global Uptime Checks: 30+ global probe locations for HTTP, HTTPS, and SSL/TLS certificate expiry detection, proactively identifying service outages.

Product Ecosystem

Cloud Monitoring

Cloud Monitoring is the core service that automatically collects GCP resource metrics (CPU, memory, network, etc.) into a time-series database. Custom metrics can be written via the Monitoring API or OpenTelemetry. Dashboards support line charts, heatmaps, and distribution charts. Service Monitoring auto-discovers microservice topologies and provides SLA/SLO/SLI management.

Cloud Logging

Cloud Logging provides centralized log collection, storage, and real-time streaming analysis for GCP resources and applications. Log Router enables flexible log distribution to BigQuery, Cloud Storage, and Pub/Sub for further processing. The Logs Explorer provides query-based log search and visualization. JSON and structured log formats are auto-parsed.

Error Reporting

Error Reporting automatically aggregates and analyzes application errors generated in GCP services. It supports error group deduplication, stack trace classification, and trend analysis. Error events can trigger Cloud Monitoring alerts and link directly to associated Cloud Logging entries. Supports Java, Python, Node.js, Go, Ruby, PHP, and .NET error format parsing.

Managed Prometheus

Managed Prometheus is GCP's native managed Prometheus service, fully compatible with PromQL and Prometheus metric formats. It supports automatic target discovery in GKE clusters without managing a Prometheus server. Collected metrics are queryable in Cloud Monitoring and can be used in Grafana dashboards. Ideal for Prometheus-standard containerized environments.

Uptime Checks

Uptime Checks provide availability monitoring from 30+ global probe locations, supporting HTTP/HTTPS endpoints, SSL certificate expiry detection, and content validation (response body keyword matching). Check results are automatically written to Cloud Monitoring metrics and can trigger alerting policies. Check frequency is configurable from 1 to 15 minutes.

Limitations

• Non-GCP Resource Limitations: Hybrid cloud requires ops-agent or Managed Prometheus collector installation, with higher configuration complexity than AWS CloudWatch Agent.
• MQL Learning Curve: Advanced queries and dashboard configuration require understanding Monitoring Query Language (MQL). PromQL users can avoid this via Managed Prometheus.
• Cross-Project View Complexity: Metrics Scope configuration for cross-project and cross-organization views involves multiple setup steps for large multi-project GCP environments.
• Modest Free Tier: Free tier has limited metric and log quotas. Beyond that, detailed Logging storage and Log Router egress costs need attention.

Use Cases

• GCP-Only Users (★★★★★): Cloud Monitoring is the foundational monitoring solution for GCP users, deeply integrated with Cloud Console and IAM. See CDN Log Analysis & Monitoring.
• Prometheus-Standard Teams (★★★★★): Managed Prometheus provides zero-ops Prometheus with native GCP integration.
• GKE Containerized Applications (★★★★☆): GKE pre-integration with Cloud Monitoring and Managed Prometheus offers an out-of-the-box Kubernetes monitoring experience.
• AI/ML Monitoring Teams (★★★★☆): Vertex AI integration supports model performance and feature monitoring for MLOps scenarios.
• Hybrid Cloud Environments (★★★☆☆): Non-GCP resources require ops-agent or third-party integrations. Compare with third-party monitoring solutions.

Pricing

Note: Google Cloud Monitoring costs are included in the unified GCP bill. Enable Cloud Billing Budgets to set budget alerts.

FAQ

• What's the relationship between Cloud Monitoring and Stackdriver? Stackdriver was the original brand name. Google acquired Stackdriver in 2016 and gradually integrated it into Cloud Monitoring and Cloud Logging. The Stackdriver brand is deprecated. See cloud monitoring services comparison.
• Can Cloud Monitor non-GCP resources? Yes. Install the ops-agent or Managed Prometheus collector on non-GCP hosts. Feature completeness is lower than GCP-native resources. See cloud monitoring services comparison.
• Cloud Monitoring vs Datadog? Cloud Monitoring: zero-cost GCP baseline with Managed Prometheus integration. Datadog: unified multi-cloud experience with richer enterprise integrations. Choose Cloud Monitoring for GCP-centric stacks, Datadog for multi-cloud. See cloud monitoring services comparison.
• Managed Prometheus vs self-hosted Prometheus? Managed Prometheus eliminates Prometheus server operations (HA, storage, scaling). Metrics are stored in Cloud Monitoring's backend with PromQL support. Migration cost is minimal for existing Prometheus users. See Prometheus monitoring setup guide.
• Which languages does Error Reporting support? Error Reporting supports error stack parsing for Java, Python, Node.js, Go, Ruby, PHP, and .NET. See Sentry error monitoring setup.