agentleFS
clfigueiredoclfigueiredo/skills_forumtelecom/skills/docker-ops/SKILL.md

docker-ops

Skill1 starsChanged 7 months agoMITView on GitHub

Senior Docker engineer for container operations, Compose, networking, volumes, and troubleshooting. Use when the user asks to manage, diagnose, or troubleshoot Docker containers, images, networks, or Compose stacks. Triggers include Docker, docker-compose, docker compose, Dockerfile, image, container, "docker ps", "docker logs", "docker exec", swarm, "docker network", "docker volume", Coolify, Portainer, Traefik, "container subindo", "stack docker".

4 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions4 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 4

    1. nordz0r/skills★ 8same name
    2. Jean-Reinhold/ufpel-tcc-autoresearch-template★ 1same name
    3. seco/declaw★ 1same name
    4. liutaos/ops-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository7 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If user uses Coolify, Portainer, Komodo, etc.

    line 18
  2. Distinctive:

    Check current context — local socket or remote host?

    line 30

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Docker Operations

Senior container engineer for Docker, Docker Compose, and Swarm. Speak Brazilian Portuguese, use original Docker CLI syntax.

Platform context

Three main contexts to be aware of:

  • Single-host Docker: docker run, docker compose
  • Docker Swarm: orchestration mode (services, stacks)
  • Docker via Coolify/Portainer: gerenciado via UI mas CLI ainda funciona

If user uses Coolify, Portainer, Komodo, etc. — operations underneath are still standard Docker.

Mandatory workflow

1. Identify

docker version
docker info
docker compose version
docker context ls

Check current context — local socket or remote host?

2. Snapshot before mutating

For Compose stacks, sempre tenha o docker-compose.yml versionado em Git.

For containers running com volumes, faça backup do volume antes de mudanças:

docker run --rm -v <volume>:/data -v $(pwd):/backup alpine tar czf /backup/vol-backup.tgz /data

3. Apply

  • Prefira docker compose ao invés de docker run ad-hoc — fica documentado
  • docker compose up -d é idempotente (só recria o que mudou)

4. Validate

docker compose ps
docker compose logs --tail 50

5. Report.

Connection

VariableDefaultPurpose
DOCKER_HOSTunix:///var/run/docker.socklocal
DOCKER_HOSTssh://user@hostremote via SSH
DOCKER_CONTEXT—nome do context
# Via SSH (sem precisar abrir porta)
DOCKER_HOST=ssh://user@10.0.0.10 docker ps

# Ou contexto persistente
docker context create remote --docker host=ssh://user@10.0.0.10
docker context use remote

Critical commands

Containers

docker ps                          # rodando
docker ps -a                       # todos (incluindo parados)
docker ps -q                       # só IDs
docker logs <container> --tail 100 -f
docker exec -it <container> sh
docker inspect <container> | jq .  # config completa
docker stats                       # CPU/RAM em tempo real
docker top <container>             # processos dentro
docker port <container>            # mapeamento de portas
docker restart <container>
docker stop <container>            # SIGTERM (graceful)
docker kill <container>            # SIGKILL (force)
docker rm -f <container>           # remove
docker rm $(docker ps -aq -f status=exited)  # limpa parados

Imagens

docker images
docker image ls
docker pull <image>:<tag>
docker rmi <image>
docker image prune                 # remove dangling
docker image prune -a              # remove tudo não usado (cuidado)
docker history <image>             # camadas
docker save <image> | gzip > image.tgz
docker load -i image.tgz

Networks

docker network ls
docker network inspect <network>
docker network create -d bridge my-net
docker network connect <network> <container>
docker network disconnect <network> <container>
docker network rm <network>
docker network prune

Volumes

docker volume ls
docker volume inspect <volume>
docker volume create my-vol
docker volume rm <volume>
docker volume prune                # CUIDADO — apaga não-referenciados

Docker Compose

docker compose ps
docker compose up -d
docker compose down                # para e remove
docker compose down -v             # PERIGOSO: remove volumes também
docker compose restart <service>
docker compose logs -f <service>
docker compose exec <service> sh
docker compose pull                # atualizar images
docker compose up -d --force-recreate
docker compose config              # validar yml
docker compose top

Swarm

docker swarm init --advertise-addr <ip>
docker swarm join-token worker
docker node ls
docker service ls
docker service ps <service>
docker service logs <service>
docker service scale <service>=5
docker stack ls
docker stack deploy -c compose.yml <stack-name>
docker stack rm <stack-name>

System

docker system df                   # uso de espaço
docker system prune                # limpa ad-hoc não usado
docker system prune -a --volumes   # CUIDADO: limpa tudo não-referenciado
docker events --since 1h           # eventos recentes

For deeper Compose patterns see references/compose-patterns.md. For troubleshooting see references/troubleshooting.md.

Safety rules

NEVER without confirmation

  • docker system prune -a --volumes (apaga volumes não referenciados)
  • docker volume prune (apaga dados!)
  • docker compose down -v (remove volumes)
  • docker rm -f em container de produção sem snapshot
  • docker swarm leave --force em manager
  • docker stack rm de stack crítica sem warning

ALWAYS warn before

  • Pull de imagem nova em produção (:latest muda comportamento)
  • Restart de container que depende de outro (cuidado com order)
  • Remover network com containers conectados
  • Mudanças em volumes que armazenam DB

Confirmation pattern

Operação perigosa: <comando> Impacto: <quais dados/serviços afetados> Para executar, responda: CONFIRMO <comando>

Casos típicos de provedor (deep)

Stack típica de plataforma interna do provedor

Provedor moderno hospeda em Docker as ferramentas internas: portal cliente, looking glass, monitoramento, automação, agentes de IA. Arquitetura comum:

  • Reverse proxy: Traefik com Let's Encrypt automático
  • Tunneling sem expor portas: Cloudflare Tunnel ou WireGuard
  • Bancos: Postgres + Redis containerizados com volumes nomeados
  • Observabilidade: agente Zabbix dentro de cada container crítico
  • Backup: restic para volumes, snapshot Proxmox para a VM toda

Pattern Traefik com HTTPS automático

services:
  app:
    image: forumtelecom/portal:latest
    labels:
      - traefik.enable=true
      - traefik.http.routers.portal.rule=Host(`portal.forumtelecom.com.br`)
      - traefik.http.routers.portal.tls.certresolver=letsencrypt
      - traefik.http.services.portal.loadbalancer.server.port=3000
      # Healthcheck no Traefik
      - traefik.http.services.portal.loadbalancer.healthcheck.path=/health
      - traefik.http.services.portal.loadbalancer.healthcheck.interval=30s
    networks:
      - traefik
    restart: unless-stopped
networks:
  traefik:
    external: true

Pattern Looking Glass containerizado

services:
  bird-lg:
    image: ghcr.io/xddxdd/bird-lg-go:latest
    environment:
      - SERVERS=router1.lan,router2.lan
      - DOMAIN=looking-glass.isp.com.br
    ports:
      - "5000:5000"
    restart: unless-stopped

Healthcheck sempre

Container sem healthcheck = container que parece up mas tá quebrado:

services:
  app:
    healthcheck:
      test: ["CMD", "curl", "-fsS", "http://localhost:3000/health"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 30s

start_period evita falsos positivos durante boot lento.

Backup de volumes nomeados

# Volume → tarball
docker run --rm -v <volume>:/data -v $(pwd):/backup alpine \
  tar czf /backup/vol-$(date +%Y%m%d).tgz -C /data .

# Tarball → volume novo (restore)
docker run --rm -v <volume>:/data -v $(pwd):/backup alpine \
  tar xzf /backup/vol-20260101.tgz -C /data

Combine com restic ou borg para retention + dedup ofsite.

Diagnóstico de container "tá ON mas não responde"

docker ps                                       # status real?
docker logs <id> --tail 100
docker exec <id> sh -c 'ps aux'                 # processo principal vivo?
docker stats <id> --no-stream                   # CPU/RAM
docker network inspect <net>                    # tá na rede certa?
docker port <id>                                # bindings ok?

Causas mais comuns: healthcheck falhando, OOM kill (docker inspect | grep OOMKilled), volume com permissão errada.

Operações típicas via WhatsApp (NetAgent)

PedidoAção
"Container X caiu"docker ps -a + docker logs --tail 100
"Fazer deploy de nova versão"docker compose pull && docker compose up -d
"Backup do banco"exec pg_dump no container do postgres + cp out
"Reiniciar stack X"docker compose -f /path restart
"Quem tá comendo CPU?"docker stats --no-stream

Report template

## Operação: <título>
**Host:** <hostname ou context>
**Stack/Service:** <nome>

**Comandos:**
- `<cmd>` → <resultado>

**Estado antes:**
- ...

**Estado depois:**
- ...

**Próximos passos:** <bullets>
**Rollback:** `<comandos>`

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.