{"status":"operational","checkedAt":"2026-10-07T10:40:11.844Z","uptime":{"24h":null,"30d":null,"90d":1},"samples":{"24h":0,"30d":0,"90d":2},"latencyMs":{"24h":null,"30d":null},"sla":{"target":0.995,"window":"monthly","cadence":"daily","credits":false,"measurement":"A scheduled probe requests a representative endpoint from each component and records the outcome. A run counts as a failure if any component fails. Figures are computed from those runs over trailing UTC days; a window with no runs reports null, not 100%.\n\nThe schedule is daily, which is the most a Vercel Hobby account allows, so each window holds roughly one sample per day. Read the percentages with the `samples` count beside them: thirty samples cannot tell 99.5% from 100%, and a single sample is a coin with one side showing. The live component checks at the top of the report are exact; only the trailing figures are sampled.","meetingTarget":null},"components":[{"name":"REST API","status":"operational","latencyMs":5732},{"name":"Scripture","status":"operational","latencyMs":6053},{"name":"Search","status":"operational","latencyMs":9721},{"name":"Cross references","status":"operational","latencyMs":9608},{"name":"MCP server","status":"operational","latencyMs":310}],"datasets":{"interlinear":"available","connections":"available"},"data":{"generatedAt":"2026-08-20T07:25:46.714Z","version":"1.6.0"},"history":{"available":true,"lastRunAt":"2026-08-23T11:12:00.497Z"}}