Steps to replicate the issue (include links if applicable):
- Execute any SPARQL query either via the frontend (https://query.wikidata.org) or via the API-backend that wasn't cached previously. This can appears to affect any query, simple or complex.
What happens?:
Since 1 or 2 days (2026-05-07 or 2026-05-07) even very simple queries time out or get a bad gateway (HTTP 502) response.
What should have happened instead?:
Get a response without 502-error and no timeouts on simple queries.
