health.live

Is this process alive: the cheapest true statement this executable can make about itself, with the version that answered. No filesystem traversal, no index build, no network — this is what a hosting platform polls, and it must cost nothing to say.

query
behaviorally_verified
module health
#health
#deployment

Exposure

Where this one definition is reachable. Absence is explicit: a surface not listed does not carry it.

surface	as
MCP	—
HTTP	GET /api/v1/live operationId health.live
command line	—

Schemas

The canonical input and output, from the Rust types; the MCP tool schema and the OpenAPI parameters and responses are derived from these.

input · Empty

No input.

No properties: the capability takes no input.

JSON Schema

{
  "additionalProperties": false,
  "description": "No input.",
  "title": "Empty",
  "type": "object"
}

output · Liveness

The answer to "is this process alive": the cheapest true thing this executable can say about itself. Deliberately not a summary of anything — a probe that grew an opinion would stop reporting the one fact the platform polls it for.

property	type	required	description
alive	boolean	yes	Always `true`. A process that could not answer would not answer at all, which is the signal.
commit	string	yes	The commit this executable was built from, in full, or `unknown` when the build could not name one. A version is what a build calls itself and two revisions share it; this is what a deployment is checked against. Never omitted: an absent field would read as a server too old to say, and `unknown` is a different, truthful answer.
dirty	boolean \| null	no	Whether that commit's tree carried uncommitted changes when this executable was built; `null` when the build did not know.
version	string	yes	This executable's version, so a rolling deployment can tell which build answered.

JSON Schema

{
  "description": "The answer to \"is this process alive\": the cheapest true thing this executable can\nsay about itself. Deliberately not a summary of anything — a probe that grew an\nopinion would stop reporting the one fact the platform polls it for.",
  "properties": {
    "alive": {
      "description": "Always `true`. A process that could not answer would not answer at all, which is\nthe signal.",
      "type": "boolean"
    },
    "commit": {
      "description": "The commit this executable was built from, in full, or `unknown` when the build could\nnot name one. A version is what a build calls itself and two revisions share it; this\nis what a deployment is checked against. Never omitted: an absent field would read as\na server too old to say, and `unknown` is a different, truthful answer.",
      "type": "string"
    },
    "dirty": {
      "description": "Whether that commit's tree carried uncommitted changes when this executable was built;\n`null` when the build did not know.",
      "type": [
        "boolean",
        "null"
      ]
    },
    "version": {
      "description": "This executable's version, so a rolling deployment can tell which build answered.",
      "type": "string"
    }
  },
  "required": [
    "alive",
    "version",
    "commit"
  ],
  "title": "Liveness",
  "type": "object"
}

Policies

benchmark: required — a target on every transport the exposure declares; the cases are the input type's
cache: disabled — every call runs the handler

Benchmark targets

Derived from the registry for this repository: one requirement per transport, and the cases the input type provides. The whole matrix is on the benchmarks page.

transport	state	cases	targets
direct	covered	1	default
http	covered	1	default