Recurring DEADLINE_EXCEEDED after 700s on simple near_text queries - Shared Cluster GCP Europe

Support
, ,
1

Environment:

  • Weaviate Cloud version: 1.34.5
  • Cloud provider: GCP
  • Region: Europe
  • Cluster type: Shared (99.5% SLA)
  • Python client: weaviate-client==4.16.9
  • Python: 3.9

Problem:

Simple near_text() queries hang for 700+ seconds before failing with DEADLINE_EXCEEDED. This is happening intermittently on our production cluster.

Timeline of incident (2025-12-30):

15:41:58 UTC - near_text() query started
[NO RESPONSE FOR 12 MINUTES]
15:53:38 UTC - DEADLINE_EXCEEDED error

Python code used:

Simple vector search - should take <100ms

response = await collection.query.near_text(
query=“CONSEILLER CLIENTÈLE BANCAIRE”,
limit=10,
return_metadata=MetadataQuery(distance=True)
)

Simple count - also hangs 700s

result = await collection.aggregate.over_all(total_count=True)

Error log:

grpc.aio._call.AioRpcError: <AioRpcError of RPC that terminated with:
status = StatusCode.DEADLINE_EXCEEDED
details = “Deadline Exceeded”
debug_error_string = “UNKNOWN:Error received from peer {grpc_status:4, grpc_message:“Deadline Exceeded”}”

weaviate.exceptions.WeaviateQueryError: Query call with protocol GRPC search failed with message Deadline Exceeded.

After the initial failure, subsequent queries fail with:
status = StatusCode.UNAVAILABLE
details = “recvmsg:Connection timed out”

Collection sizes:

  • StaticRole: ~few thousand objects
  • Skill: ~18,800 objects
  • JobTitle: ~few thousand objects

Questions:

  1. Is this expected behavior on Shared tier during high-traffic periods?
  2. Are there known issues with GCP Europe region or v1.34.5?
  3. What timeout configuration do you recommend for Shared clusters?
  4. Should we implement specific retry/circuit-breaker patterns?
2

Hi @Mehdi_Essafi

I’ve received your ticket on support@weaviate.io and have responded to you there.

As you’re a customer on our cloud platform, please use support@weaviate.io for future support requests. This will allow you to open a ticket in our system and access your cloud clusters more efficiently, ensuring a faster resolution.

Best regards,
Mohamed Shahin
Weaviate Support Engineer