User Details
User Details
- User Since
- Nov 2 2020, 1:15 PM (300 w, 4 d)
- Availability
- Available
- IRC Nick
- gmodena
- LDAP User
- Gmodena
- MediaWiki User
- GModena (WMF) [ Global Accounts ]
Fri, Jul 24
Fri, Jul 24
gmodena changed the status of T432775: Improve WDQS v1 telemetry with information about payload size. from Open to In Progress.
gmodena changed the status of T432775: Improve WDQS v1 telemetry with information about payload size., a subtask of T432467: [PP3.5.1] WDQS Cost-Based Rate Limiting, from Open to In Progress.
gmodena added a comment to T428631: wdqs-proxy: implement custom envoy latency bucket reporting.
Pinning this to PP3.5.1.
gmodena updated the task description for T421984: Add grafana panels and alerts for wdqs autoremediation metrics..
Thu, Jul 23
Thu, Jul 23
gmodena updated the task description for T428235: data-reload: create a QLever index using an ephemeral instance.
gmodena placed T428235: data-reload: create a QLever index using an ephemeral instance up for grabs.
gmodena renamed T428235: data-reload: create a QLever index using an ephemeral instance from data-reload: create a QLever index using an ephemeral instance to [NEED REGROOMING] data-reload: create a QLever index using an ephemeral instance.
gmodena added a comment to T428235: data-reload: create a QLever index using an ephemeral instance.
Thanks for tihs!
gmodena updated the task description for T431395: [SPIKE][NEEDS GROOMING] resource allocation for co-located wdqs-qlever instances.
Wed, Jul 22
Wed, Jul 22
gmodena updated the task description for T431395: [SPIKE][NEEDS GROOMING] resource allocation for co-located wdqs-qlever instances.
Tue, Jul 21
Tue, Jul 21
Mon, Jul 20
Mon, Jul 20
gmodena added a comment to T405949: WDQS sparql query event generator should not set meta.dt.
gmodena updated the task description for T432627: wdqs-qlever: deploy and test QLevers index hot swap patch..
gmodena placed T432348: wdqs v2: deploy wdqs-next up for grabs.
Fri, Jul 17
Fri, Jul 17
gmodena added a comment to T432440: Enable persisting updates on Qlever.
Thanks for this!
Thu, Jul 16
Thu, Jul 16
gmodena added a comment to T430782: consumer: support for large batch sizes..
Wed, Jul 15
Wed, Jul 15
gmodena added a comment to T313813: API Gateway to provide authorization and capacity management for W[CD]QS.
gmodena renamed T431976: wdqs v2: Setup internal routing for query federation between main and scholarly from wdqs v2: test external and internal federation to [NEEDS GROOMING] wdqs v2: test external and internal federation.
gmodena changed the status of T422179: Set up regular bulk data ingestion and indexing pipeline, a subtask of T431253: [DE2.4.2] WDQS v2 Scaling, from Open to Stalled.
gmodena changed the status of T422179: Set up regular bulk data ingestion and indexing pipeline from Open to Stalled.
Tue, Jul 14
Tue, Jul 14
gmodena updated subscribers of T430782: consumer: support for large batch sizes..
Mon, Jul 13
Mon, Jul 13
gmodena renamed T421200: wdqs-proxy: event platform integration for query logs from [NEEDS GROOMING] wdqs-proxy: event platform integration for query logs to wdqs-proxy: event platform integration for query logs.
Fri, Jul 10
Fri, Jul 10
gmodena moved T429150: WDQS v2 subdomain for public endpoints from Backlog to In Engineering on the Wikidata Platform Team (Sprint 07 (2026/07/07) board.
gmodena updated the task description for T427785: wdqs-proxy: Allow the usage of the Qlever pathSearch SERVICE.
gmodena moved T431310: wdqs-proxy: disable query rewrite code path from Backlog to Ready on the Wikidata Platform Team (Sprint 07 (2026/07/07) board.
Thu, Jul 9
Thu, Jul 9
gmodena updated the task description for T431690: [NEEDS GROOMING][SPIKE] what is required to enable distributed tracing on WDQSv2?.
gmodena added a comment to T431271: [SPIKE] understand, document and tune QLever memory behaviour when real time updates are applied..
gmodena updated the task description for T431052: consumer: collect QLever memory allocation metrics with different batch sizes.
Jul 8 2026
Jul 8 2026
gmodena renamed T431390: We need the capability to build and publish qlever from source from [NEEDS GROOMING] We need the capability to build and publish qlever from source to We need the capability to build and publish qlever from source.
gmodena updated the task description for T431390: We need the capability to build and publish qlever from source.
gmodena added a comment to T431531: Request creation of wdp-gitlab-runner VPS project.
Terrific! Thank you so much for picking this task up so quickly. Now let's build :)
gmodena added a comment to T431531: Request creation of wdp-gitlab-runner VPS project.
gmodena added a comment to T431438: discovery.wikibase_rdf_subgraphs (/wmf/data/discovery/wikidata/rdf_subgraphs) is using 34Tb and is not properly cleaned up.
@dcausse terrific. Thanks!
gmodena updated the task description for T422179: Set up regular bulk data ingestion and indexing pipeline.
gmodena added a comment to T422179: Set up regular bulk data ingestion and indexing pipeline.
Keeping this "in engineering" to account for post-merge and deploy fixes for the pipeline.
gmodena added a comment to T431531: Request creation of wdp-gitlab-runner VPS project.
AFAIK those are community driven hosting projects that run their own query service.
gmodena added a project to T431531: Request creation of wdp-gitlab-runner VPS project: Wikidata Platform Team.
gmodena renamed T431531: Request creation of wdp-gitlab-runner VPS project from Request creation of wdqs-qlever-build-runner VPS project to Request creation of wdp-gitlab-runner VPS project.
gmodena renamed T431390: We need the capability to build and publish qlever from source from We need the capability to build and publish qlever from source to [NEEDS GROOMING] We need the capability to build and publish qlever from source.
Jul 7 2026
Jul 7 2026
gmodena added a comment to T431390: We need the capability to build and publish qlever from source.
Discussed internally in: https://wikimedia.slack.com/archives/C05H0JYT85V/p1783414337204559
gmodena changed the point value for T428235: data-reload: create a QLever index using an ephemeral instance from 1 to 2.
gmodena renamed T431445: wdqs-streaming-consumer: gracefully handle update failures from wdqs-streaming-consumer: Gracefully handle update failures to wdqs-streaming-consumer: gracefully handle update failures.
gmodena moved T429250: data-reload: implement gen_rev_map task from Backlog to In Code Review on the Wikidata Platform Team (Sprint 07 (2026/07/07) board.
gmodena moved T429252: data-relaod: implement split_subgraphs task from Backlog to In Code Review on the Wikidata Platform Team (Sprint 07 (2026/07/07) board.
gmodena edited projects for T421200: wdqs-proxy: event platform integration for query logs, added: Wikidata Platform Team (Sprint 07 (2026/07/07); removed Wikidata Platform Team (Sprint 06 (2026/06/02)).
gmodena edited projects for T422081: Create a migration progress dashboard, added: Wikidata Platform Team (Sprint 07 (2026/07/07); removed Wikidata Platform Team (Sprint 06 (2026/06/02)).
gmodena edited projects for T429150: WDQS v2 subdomain for public endpoints, added: Wikidata Platform Team (Sprint 07 (2026/07/07); removed Wikidata Platform Team (Sprint 06 (2026/06/02)).
gmodena added a comment to T431438: discovery.wikibase_rdf_subgraphs (/wmf/data/discovery/wikidata/rdf_subgraphs) is using 34Tb and is not properly cleaned up.
Thanks @dcausse we plan to use use the same spark pipeline to write into a wikidata db. The content in discovery.wikibase_rdf_subgraphs should be safe to be deleted, but we'll hit the same issue with data retention (the plan was to use the same purging mechanism https://gitlab.wikimedia.org/repos/data-engineering/airflow-dags/-/merge_requests/2321).
gmodena added a comment to T425119: data-reload: implement a quality check step on the wikidata dump.
@trueg I'm afraid my collab notes got crossed, bear with me. IIRC we are considering treating the ttl file as temp data, to be deleted (together with the munged temp files) when the pipeline completes.
If we do this, we would break historical comparisons.
gmodena added a comment to T422038: WDQS: write puppet code/investigate performance optimizations for new hardware.
Had a chat about HW with the QLever team. Their experience is usually no OS tuning is required. They run their own endpoint (qlever.dev) on a vanilla ubuntu setup
gmodena placed T431310: wdqs-proxy: disable query rewrite code path up for grabs.
gmodena added a parent task for T428960: WDQS on K8s: test common operations/failure scenarios: T431253: [DE2.4.2] WDQS v2 Scaling.
gmodena updated the task description for T431392: [SPIKE] evaluate TopoLVM viability as the storage for the indexing pipeline.