Shared log readiness

When to use this

Use this pattern when several services print the same startup signal. Instead of repeating the same [process.readiness] block everywhere, define it once under [global.readiness].

Example

[global]
log_mode = "both"
log_format = "json"

[global.readiness]
type = "log"

[global.readiness.ready]
stream = "stdout"
pattern = "service_ready"

[[process]]
name = "billing-public"
command = "./tools/run_service.sh billing-public"

[[process]]
name = "billing-worker-rollup"
command = "./tools/run_service.sh billing-worker-rollup"

[[process]]
name = "billing-grpc"
command = "./tools/run_service.sh billing-grpc"

Why this works

Each process inherits global.readiness. This keeps large service lists short and makes the contract explicit: a service is ready when it prints service_ready on stdout.

This works best when all services are wrapped by a common launcher or framework that emits the same startup event.

Defaults involved here

• process.readiness: inherits global.readiness
• ready.stream: default stdout, shown explicitly here for clarity
• instances: default 1
• show_timestamps: effective default true
• log_dir: default .vectron/logs

What to customize

• Replace service_ready with the exact ready event emitted by your services.
• Use a process-level readiness override for services that need a different signal.
• Keep the pattern narrow enough that normal log lines cannot accidentally mark a service ready.

See also

• Readiness probes
• Large service workspace
• vectron.toml reference