AC:
- Add a new field to record (raw) response size (in bytes) in WDQS/Blazegraph telemetry.
AC:
| Subject | Author | Repo | Branch | Lines +/- | |
|---|---|---|---|---|---|
| deploy version 0.3.165 | Ryan Kemper | wikidata/query/deploy | master | +15 -15 | |
| telemetry: log response size in bytes for query events | Gmodena | wikidata/query/rdf | master | +227 -24 |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | BTracy-WMF | T432467 [PP3.5.1] WDQS Cost-Based Rate Limiting | |||
| Resolved | gmodena | T432775 Improve WDQS v1 telemetry with information about payload size. |
gmodena opened https://gitlab.wikimedia.org/repos/data-engineering/schemas-event-secondary/-/merge_requests/132
sparql/query: add response_size field
gmodena updated https://gitlab.wikimedia.org/repos/data-engineering/schemas-event-secondary/-/merge_requests/132
sparql/query: add response_size field
Change #1313342 had a related patch set uploaded (by Gmodena; author: Gmodena):
[wikidata/query/rdf@master] telemetry: log response size in bytes for query events
gmodena merged https://gitlab.wikimedia.org/repos/data-engineering/schemas-event-secondary/-/merge_requests/132
sparql/query: add response_size field
Change #1313342 merged by jenkins-bot:
[wikidata/query/rdf@master] telemetry: log response size in bytes for query events
Change #1323142 had a related patch set uploaded (by Ryan Kemper; author: Ryan Kemper):
[wikidata/query/deploy@master] deploy version 0.3.165
Change #1323142 merged by Ryan Kemper:
[wikidata/query/deploy@master] deploy version 0.3.165
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:16:33Z] <ryankemper@deploy1003> Started deploy [wdqs/wdqs@cf64fe9]: T432775: 0.3.165 canary (response_size telemetry)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:17:01Z] <ryankemper@deploy1003> Finished deploy [wdqs/wdqs@cf64fe9]: T432775: 0.3.165 canary (response_size telemetry) (duration: 00m 28s)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:18:52Z] <ryankemper@deploy1003> Started deploy [wdqs/wdqs@cf64fe9]: T432775: 0.3.165 canary (response_size telemetry)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:18:59Z] <ryankemper@deploy1003> Finished deploy [wdqs/wdqs@cf64fe9]: T432775: 0.3.165 canary (response_size telemetry) (duration: 00m 06s)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:28:19Z] <ryankemper@deploy1003> Started deploy [wdqs/wdqs@cf64fe9]: T432775: deploy 0.3.165 (response_size telemetry)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T04:47:42Z] <ryankemper@deploy1003> Finished deploy [wdqs/wdqs@cf64fe9]: T432775: deploy 0.3.165 (response_size telemetry) (duration: 19m 22s)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T05:28:28Z] <ryankemper@deploy1003> Started deploy [wdqs/wdqs@cf64fe9] (wcqs): T432775: deploy 0.3.165 (response_size telemetry)
Mentioned in SAL (#wikimedia-operations) [2026-08-08T05:31:04Z] <ryankemper@deploy1003> Finished deploy [wdqs/wdqs@cf64fe9] (wcqs): T432775: deploy 0.3.165 (response_size telemetry) (duration: 02m 36s)
Deployed 0.3.165 to prod 2026-08-07 evening
wdqs-external, wdqs-internal, and wcqs-external sparql-query streams all show http.response_size getting populated now, and I ran a few test curls that came through the other end with the correct byte size
Note that there will be some events on 1.4.0 schema that don't have http.response_size because the event platform occasionally emits synthetic canary events into every stream to verify liveness. Practically speaking this means when looking at the data we want to filter on presence of http.response_size rather than just keying on the schema value