Hi everyone,
Yesterday we ran a small bulk import (~2k documents) on some of our Firestore MongoDB Connector instances. The import itself completed fine, but shortly after we started getting intermittent 504: Deadline expired before operation could complete errors.
The most puzzling part: a completely separate instance, which had no import at all and contained only a few dozen records, was also affected. Even a simple find() on a tiny collection was returning 504.
What made it particularly confusing:
- Local connections with the same connection string worked fine
- The Google Cloud Console UI was also intermittently failing
- Google Service Health showed no active incidents
- It was not a consistent failure - sometimes queries went through, sometimes they didn’t
This strongly suggests the bulk import caused some kind of resource pressure or throttling at the project level, spilling over to all connector instances regardless of their data size or whether they were involved in the import.
Has anyone experienced something similar? Any idea what could cause this? Is there a known interaction between bulk imports and overall connector stability?
Thanks