diff options
| author | Ergün Erdoğmuş <erdogmusergun@gmail.com> | 2026-11-12 11:05:46 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2026-11-12 11:05:46 +0100 |
| commit | a4830052c795edbce8f991debc69a24d2171785a (patch) | |
| tree | 5177757c6df8284f720443c83fdd88265fde8c8e | |
| parent | 1a2c79d4e6c4a2283dc2c614b6538cbe1ed5ca62 (diff) | |
vercel-optimize: document queryId and meta in vercel metrics JSONvercel-optimize-metrics-json-meta
Co-authored-by: vercel-v[bot] <353008837+vercel-v[bot]@users.noreply.github.com>
| -rw-r--r-- | skills/vercel-optimize/references/data-collection.md | 12 |
1 files changed, 11 insertions, 1 deletions
diff --git a/skills/vercel-optimize/references/data-collection.md b/skills/vercel-optimize/references/data-collection.md index 140c787..93bcb17 100644 --- a/skills/vercel-optimize/references/data-collection.md +++ b/skills/vercel-optimize/references/data-collection.md @@ -154,12 +154,22 @@ The skill never crashes the entire collection on a single endpoint failure. Ever { "timestamp": "2026-04-13T04:00:00.000Z", "vercel_request_count_sum": 0, "route": "/dashboard/[sessionId]" } /* ... */ ], - "statistics": { "bytesRead": 10267, "rowsRead": 947, "dbTimeSeconds": 0 } + "statistics": { "bytesRead": 10267, "rowsRead": 947, "dbTimeSeconds": 0 }, + // Newer CLIs (with `vercel metrics --data`) also return: + "queryId": "observability-api-...", + "meta": { + "outputs": { + "vercel_request_count_sum": { "kind": "metric", "metric": "vercel.request.count", "unit": "count" } + }, + "effectiveTimeRange": { "start": "2026-04-13T04:00:00.000Z", "end": "2026-05-13T08:00:00.000Z" } + } } ``` Field naming rule: the metric ID's dots become underscores, and the aggregation suffix is appended — `vercel.request.count` + `sum` → `vercel_request_count_sum`. `lib/vercel.mjs::normalizeSummary()` flattens `summary[]` into `[{<dim>: v, ..., value: <n>}]`. +`queryId` and `meta` are additive, so the collector ignores them. On those CLIs, `query.startTime` and `query.endTime` report the range the API actually queried (`meta.effectiveTimeRange`, after bucket alignment), which can be wider than the requested `--since` window. `meta.outputs` is keyed by the same `<metric>_<aggregation>` column, and `meta.excludedIntervals` lists Observability Plus periods left out of the result. `vercel metrics --data` returns the same `query`/`summary`/`data` layout with one column per output alias, for queries with several metrics or formulas. + ### `vercel metrics schema --format json` Array of `{id, description}` entries — NOT an object. Many metric IDs in earlier docs don't exist: there is no `vercel.function.cold_starts`, no `vercel.cache.hits`. Cache state is the `cache_result` dimension on `vercel.request.count`. |
