As pointed out by jayme on #wikimedia-serviceops, the timeouts in the Termbox SSR service on Kubernetes are constantly increasing.
We should understand what causes these timeouts, specifically. That knowledge should be documented (probably at https://wikitech.wikimedia.org/wiki/WMDE/Wikidata/SSR_Service).
Further actions may or may not follow from understanding these timeouts.
What we already know:
- these timeouts also happen on smallish entities, not only on large ones