Types of logs you can collect
| Log Type | AWS | Azure | GCP |
|---|---|---|---|
| Application | CloudWatch Logs | Log Analytics | Cloud Logging |
| Platform | ECS/EKS logs | AKS/App Service | GKE/Cloud Run |
| Audit | CloudTrail | Activity Logs | Audit Logs |
| Network | VPC Flow Logs | NSG Flow Logs | VPC Flow Logs |
| Service | ALB, CloudFront | Load Balancer | Cloud CDN |
Questions you can answer
| With cloud logs in Grafana, you can answer… |
|---|
| What errors are occurring across my services? |
| Who accessed what resources and when? |
| What happened in the 5 minutes before the outage? |
| Which API calls are failing and why? |
Script
Cloud environments generate a lot of logs, way more than you might expect. Let’s break down what’s out there.
Application logs are what your code outputs: debug statements, error messages, transaction records.
Platform logs come from managed services, what’s happening in your Kubernetes cluster, your App Service, your Cloud Run containers.
Audit logs are the security goldmine: who logged in, who changed what configuration, which API calls were made. These are often required for compliance.
Network logs capture traffic patterns, VPC flow logs showing what’s talking to what, useful for security analysis and network troubleshooting.
Service logs from load balancers and CDNs show request details, latency breakdowns, error responses.
You probably don’t need all of these immediately. Start with what helps you debug production issues, usually application and platform logs. Add audit logs if you have compliance requirements. Layer in network and service logs as your observability practice matures.