A container is running but your app feels sluggish. Is it CPU-bound? Leaking memory? Stuck on a runaway process? Compose gives you two essential commands to find out: docker compose top and docker compose stats.

docker compose top

Show running processes inside each service’s container:

docker compose top

myapp-web-1
UID      PID     PPID    C    STIME   TTY   TIME       CMD
root     12345   12320   0    09:15   ?     00:00:02   nginx: master process
nobody   12400   12345   0    09:15   ?     00:00:00   nginx: worker process

myapp-api-1
UID      PID     PPID    C    STIME   TTY   TIME       CMD
node     12500   12480   2    09:15   ?     00:01:45   node server.js
node     12600   12500   0    09:15   ?     00:00:12   node worker.js

You see every process running inside each container with their CPU usage (C column) and CPU time. Perfect for spotting runaway workers or unexpected child processes.

Target a specific service if you don’t want the full list:

docker compose top api

docker compose stats

Real-time CPU, memory, network, and disk I/O for services in your Compose project, scoped automatically to the current project:

# Live view, updates every second
docker compose stats

# One-shot snapshot instead of live view
docker compose stats --no-stream

# Focus on specific services
docker compose stats api db

Output:

CONTAINER ID   NAME              CPU %    MEM USAGE / LIMIT     MEM %    NET I/O          BLOCK I/O
a1b2c3d4e5f6   myapp-api-1       45.23%   312.5MiB / 512MiB     61.04%   1.2MB / 3.4MB    0B / 0B
b2c3d4e5f6a7   myapp-web-1       0.12%    8.3MiB / 256MiB       3.24%    524kB / 1.1MB    0B / 0B
c3d4e5f6a7b8   myapp-db-1        2.50%    98.4MiB / 1GiB        9.61%    245kB / 892kB    12MB / 45MB

This tells you at a glance which service is hot. In the example above, api is using 45% CPU and near its memory limit, a good place to start investigating.

Unlike docker stats (which shows every container on the host), docker compose stats shows only the containers in your current Compose project, no need to filter manually.

Custom format

Narrow down to just the columns you care about:

docker compose stats --no-stream \
  --format "table {{.Name}}\t{{.CPUPerc}}\t{{.MemPerc}}"

NAME            CPU %    MEM %
myapp-api-1     45.23%   61.04%
myapp-web-1     0.12%    3.24%
myapp-db-1      2.50%    9.61%

Combining with resource limits

These monitoring commands pair naturally with Tip #16 (resource limits). Set limits, then watch to see if services are hitting them:

services:
  api:
    image: myapp-api
    deploy:
      resources:
        limits:
          cpus: "1.0"
          memory: 512M

# Watch the container approach its limits
docker compose stats api

A service consistently at 95%+ of its limits is a hint you need to either raise the limit or optimize the code.

Quick debugging workflow

When something feels off:

# 1. See what's running
docker compose ps

# 2. Check resource usage
docker compose stats --no-stream

# 3. Drill into the hot service
docker compose top api

# 4. Check recent events for restarts or OOM kills
docker compose events --since 5m

This four-command sequence covers 80% of “my stack feels slow” investigations.

Further reading