Platform Health

One workflow, @dataverket/platform-health, that checks the forges, S3 gateways and HTTP endpoints a repository names and fails when one of them is down. Each run calls the read-only health method of every @dataverket/forgejo model in forgeModels, every @dataverket/versitygw/gateway model in gatewayModels and every @dataverket/http/endpoint model in endpointModels, which records each one's version and health, or reachability, TLS and latency, or status, TLS, redirects and latency, as that model's own record. One assertion per model then fails the run when a forge answers unhealthy, a gateway is unreachable or an endpoint is not ok, or either is slower than maxLatencyMs when that is set. Nothing is written to any service. The models are the caller's, named by name, so URLs, tokens and CA files stay in the caller's definitions and no input ever carries a value. A run takes a few seconds and is made to be run often: by hand as the morning check, on a schedule under swamp serve, or as the first step of a larger workflow. Built on @dataverket/forgejo, @dataverket/versitygw and @dataverket/http, all three dependencies.