Məlumat bazası

Containers can silently consume excessive CPU or memory, affecting other containers and the host itself. This guide covers monitoring Docker specifically, beyond general server monitoring tools.

Quick Real-Time Stats

docker stats

Shows live CPU%, memory usage/limit, network I/O, and block I/O for every running container — the fastest way to spot a runaway container.

Stats for a Specific Container

docker stats CONTAINER_NAME

One-Time Snapshot (Non-Streaming)

docker stats --no-stream

Useful for scripting or logging periodic snapshots rather than watching a live stream.

Checking Overall Docker Disk Usage

docker system df

Shows space consumed by images, containers, volumes, and build cache — useful for identifying what's filling up /var/lib/docker.

Detailed Breakdown

docker system df -v

Checking Individual Container Logs Size

sudo du -sh /var/lib/docker/containers/*/*-json.log | sort -hr

See Managing Docker Logs if a specific container's logs are consuming excessive disk space.

Setting Resource Limits to Prevent Runaway Containers

services:
  web:
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 512M

This caps what any single container can consume, protecting the host and other containers.

Monitoring with cAdvisor (Container-Specific Metrics for Prometheus)

docker run -d \
  --name cadvisor \
  -p 8080:8080 \
  -v /:/rootfs:ro \
  -v /var/run:/var/run:ro \
  -v /sys:/sys:ro \
  -v /var/lib/docker/:/var/lib/docker:ro \
  gcr.io/cadvisor/cadvisor:latest

cAdvisor exposes detailed per-container metrics that can be scraped by Prometheus — see How to Set Up Prometheus and Grafana for VPS Monitoring for the full monitoring stack integration.

Checking Container Health Status

docker ps

The STATUS column shows (healthy)/(unhealthy) for containers with a configured healthcheck — see Docker Restart Policies and Healthchecks Explained.

Identifying a Runaway Container

docker stats --no-stream --format "table {{.Name}}\t{{.CPUPerc}}\t{{.MemUsage}}" | sort -k2 -hr

Common Errors

A container consumes far more memory than expected — likely a memory leak in the application; set a hard memory limit as a safety net while investigating the root cause.

Host running out of disk from Docker — check docker system df and clean up with docker system prune -a (review what will be removed first).

Best Practices

• Set resource limits on every production container, not just the ones currently causing problems
• Monitor docker system df periodically, not just when disk space becomes critical
• Use cAdvisor with Prometheus/Grafana for historical container metrics rather than only real-time snapshots

Related Articles

• Docker Security Best Practices for Production Servers
• Managing Docker Logs
• Docker Restart Policies and Healthchecks Explained