Probe options¶

Parameter	Description
`name`	Probe identifier (e.g. `"liveness"`, `"readiness"`, `"startup"`).
`checks`	List of health checks to run.
`summary`	Custom description for the probe (used in responses). If omitted, a default is generated from `name`.
`default_check_timeout_ms`	Reserved legacy field; it does not alter execution. Use `RunPolicy.probe_timeout_ms` for a probe-wide timeout.

To customize HTTP responses, pass options=build_probe_route_options(...) to HealthcheckRouter or health(). Build options with:

Parameter	Description
`success_handler`	Handler for healthy responses. Receives `ProbeAsgiResponse`, returns response body (or `None` for empty).
`failure_handler`	Handler for unhealthy responses. Same signature as `success_handler`.
`success_status`	HTTP status for healthy (default: `204` No Content).
`failure_status`	HTTP status for unhealthy (default: `503`).
`debug`	Include check details in responses (default: `False`).
`prefix`	URL prefix for probe routes (default: `"/health"`).
`timeout`	Max seconds for all checks; on exceed returns failure (default: `None` = no limit).

Example: HealthcheckRouter(Probe(...), options=build_probe_route_options(debug=True, prefix="/health")).

Debug mode¶

With debug=True, unhealthy responses bypass the failure handler by design and dump the full report, including each error.message ("<ExceptionType>: <message>"). Error metadata is redacted by key, but messages are free text and may name internal hosts or dependency details. Keep debug=False (the default) on endpoints reachable from outside the deployment; enable it only for internal diagnostics.

RunPolicy Options¶

When using [ProbeRunner][] directly, you can customize execution behavior with RunPolicy:

Parameter	Description
`mode`	Controls probe-timeout handling: `"strict"` raises `HealthCheckTimeoutError`; `"reporting"` returns a failed report. Ordinary failed checks remain failed in both modes. Default: `"strict"`.
`execution`	Controls check execution within one probe: `"parallel"` or `"sequential"`. Default: `"parallel"`.
`probe_timeout_ms`	Timeout in milliseconds for one `runner.run(probe)` call. Default: `None`.
`health_evaluation`	Controls evaluation strategy: `"all_required"` (all probes must pass) or `"partial_allowed"` (some probes can fail without failing overall). Default: `"all_required"`.
`max_concurrency`	Caps how many checks run at once in parallel mode. `None` removes the cap. Default: `8`.

Example:

from fast_healthchecks import ProbeRunner, RunPolicy

policy = RunPolicy(mode="strict", execution="parallel", probe_timeout_ms=5000)

async with ProbeRunner(policy) as runner:
    report = await runner.run(probe)