Operations

Operations capabilities do not directly carry business data. They provide audit, health checks, instance heartbeats, metrics, and a runtime console around production operations. They also connect through Provider and the lifecycle, so they enter the dependency graph only when installed.

If you are writing your first route, you can skip this section for now. Come back when the application is ready to deploy, needs troubleshooting, or must keep audit records.

Common production capabilities

Recommended first operations capabilities

During development, start with observe if you need /health and /ready. In production, a common first setup is:

app.Install(
    log.Provider(),
    observe.Provider(observe.Config{Service: "api", Mount: "/-"}),
)

If you need audit or cluster heartbeats, install them next:

app.Install(
    queue.Provider(),
    audit.Provider(audit.Config{Writer: audit.DefaultLogWriter()}),
    cluster.Provider(cluster.DriverWith(cluster.MemoryDriver())),
)

console is useful for protected internal environments. Do not expose it directly to the public internet without authentication and network isolation.

Should small projects use operations capabilities

Add them by stage:

• Local development: skip them, or install only observe.
• Test environment: install observe and log.
• Production: add audit for compliance, cluster for multi-instance deployments, and console only when a runtime panel is needed.

How they differ from business capabilities

Business capabilities are usually called directly by business code, such as cache.Default() or queue.Default(). Operations capabilities are more often consumed by middleware, Host units, or the console:

• Audit writes records through HTTP middleware.
• Observe exposes health checks and metrics through HTTP routes.
• Cluster starts heartbeats through the lifecycle.
• Console reads runtime information from other capabilities through panels.

Common adoption order

A practical production order is:

1. Install log first, so errors and access logs have a place to go.
2. Add observe for load balancers, container platforms, or monitoring systems.
3. Add audit when compliance or business-operation tracing is required.
4. Add cluster for multi-instance deployments.
5. Add console only when you need an internal runtime panel.

Do not install every operations package at the beginning. Start with a running application, then add what the deployment environment actually needs.