Monitoring and Logs Intro

Metrics show health over time

Metrics are numeric measurements: CPU percent, request count, error rate, disk space. Dashboards help you spot trends; alarms notify you when a threshold is crossed.

Logs tell the story

Logs are timestamped event records from apps, load balancers, and cloud APIs. Centralizing logs makes debugging and audits possible. Providers offer services such as CloudWatch, Azure Monitor, and Cloud Logging.

Start small

• Enable basic host or service metrics on anything you leave running
• Turn on audit or activity logs for the account
• Alert on billing thresholds and obvious failures (5xx spikes, instance down)
• Keep retention long enough to investigate incidents, short enough to control cost

Tail a conceptual log query

Exact commands differ; the pattern is "select recent events for a resource or severity."

# Examples — product names and filters vary
# aws logs tail /aws/lambda/my-function --follow
# az monitor activity-log list --max-events 20 --output table
# gcloud logging read 'severity>=ERROR' --limit 20 --format json