dokploy-health-patterns

Dokploy Health Patterns

When to Use This Skill

• When adding health checks to any Dokploy service
• When configuring service dependencies with service_healthy
• When troubleshooting container startup order issues
• When user asks about "health checks" or "service dependencies"

When NOT to Use This Skill

• For services that don't support health checks (rare, most do)
• For helper services that start instantly (use service_started instead)

Prerequisites

• Understanding of service internals (ports, health endpoints)
• Knowledge of available tools in container (curl, wget, etc.)

---

Core Pattern

Standard Health Check Template

Every health check follows this structure:

healthcheck:
  test: ["CMD", "command", "args"]
  interval: 30s       # How often to check
  timeout: 10s        # Max time per check
  retries: 3          # Failures before unhealthy
  start_period: 30s   # Grace period for startup

Configuration Guidelines:

Parameter	Default	When to Adjust
interval	30s	Increase for resource-constrained systems
timeout	10s	Increase for slow-starting services
retries	3	Increase for flaky services
start_period	30s	Increase for complex services (databases, Java apps)

---

Health Check Patterns by Service Type

Pattern 1: HTTP Services (curl)

For services with HTTP endpoints (most web apps):

healthcheck:
  test: ["CMD", "curl", "-f", "http://localhost:${PORT}/"]
  interval: 30s
  timeout: 10s
  retries: 3
  start_period: 30s

Variations:

# With specific health endpoint
test: ["CMD", "curl", "-f", "http://localhost:3000/health"]

# With API endpoint
test: ["CMD", "curl", "-f", "http://localhost:3000/api/healthz"]

# Silent output
test: ["CMD", "curl", "-sf", "http://localhost:8080/ping"]

Pattern 2: HTTP Services (wget - Alpine-based)

For Alpine-based images without curl:

healthcheck:
  test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://localhost:${PORT}"]
  interval: 30s
  timeout: 10s
  retries: 3
  start_period: 30s