Monitoring Resource Usage

Start reading

LEVEL 0

The Problem

Your containers are running. But are they using too much CPU? Too much memory? Is one container hogging resources?

You need to monitor resource usage in real-time to:

Identify performance bottlenecks
Detect resource leaks
Set appropriate resource limits
Plan capacity

LEVEL 1

The Concept — The Dashboard Gauges

The Concept

Imagine a car dashboard.

You have gauges for:

Speed (how fast you’re going)
RPM (engine activity)
Fuel (resource remaining)
Temperature (system health)

docker stats is the dashboard for your containers.

It shows CPU, memory, network, and disk usage in real-time.

LEVEL 2

The Mechanics — Using docker stats

The Mechanics

View all running containers:

docker stats

Output:

CONTAINER ID   NAME      CPU %    MEM USAGE / LIMIT     MEM %    NET I/O         BLOCK I/O
abc123         web       0.25%    45.5MiB / 512MiB      8.89%    1.2MB / 800kB   0B / 0B
def456         db        1.50%    250MiB / 1GiB         24.41%   500kB / 1.2MB   10MB / 5MB

View specific containers:

docker stats web db

No streaming (one-shot):

docker stats --no-stream

Shows current stats and exits (doesn’t continuously update).

Format output:

docker stats --format "table {{.Name}}\t{{.CPUPerc}}\t{{.MemUsage}}"

LEVEL 4

Setting Resource Limits

Limit memory:

services:
  app:
    image: myapp
    mem_limit: 512m           # Hard limit
    mem_reservation: 256m     # Soft limit (reservation)

Or:

docker run -m 512m myapp

Limit CPU:

services:
  app:
    image: myapp
    cpus: 1.5                 # Use up to 1.5 CPU cores

Or:

docker run --cpus=1.5 myapp

CPU shares (relative weight):

services:
  high-priority:
    image: myapp
    cpu_shares: 1024          # Higher priority

  low-priority:
    image: worker
    cpu_shares: 512           # Lower priority

Combined example:

services:
  web:
    image: nginx
    mem_limit: 256m
    cpus: 0.5
  
  app:
    image: myapp
    mem_limit: 1g
    cpus: 2
  
  db:
    image: postgres:15
    mem_limit: 2g
    cpus: 2
    mem_reservation: 1g

LEVEL 5

Monitoring in Production

For production, use proper monitoring tools:

Prometheus + Grafana:

services:
  prometheus:
    image: prom/prometheus
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml
      - prometheus-data:/prometheus
    ports:
      - "9090:9090"
  
  grafana:
    image: grafana/grafana
    ports:
      - "3000:3000"
    volumes:
      - grafana-data:/var/lib/grafana

cAdvisor (Container Advisor):

Collects resource usage metrics for all containers:

services:
  cadvisor:
    image: gcr.io/cadvisor/cadvisor:latest
    volumes:
      - /:/rootfs:ro
      - /var/run:/var/run:ro
      - /sys:/sys:ro
      - /var/lib/docker/:/var/lib/docker:ro
    ports:
      - "8080:8080"

View metrics at http://localhost:8080

Cloud monitoring:

AWS CloudWatch (for ECS)
Google Cloud Monitoring
Azure Monitor
Datadog
New Relic