Azure deployement with weaviate as sidecar container

General
1

I am trying to deploy fastapi app using weaviate service, currently i have setup code locally through docker-compose using openai embeddings, there is no option in azure sidecar container for exposing grpc_port, and that why i am getting
Is Weaviate running and reachable at http://localhost:8080?
everything is correctly configured but no way it is working, is there any way we can run weaviate without requiring grpc port? (also i think azure recently removed yaml file option for multi container and introduced as sidecar container)

attaching screenshot here with for azure side setup

image
image1609Ă—850 48.3 KB

2

Yep, you’ve run into one of the recent Azure sidecar changes: gRPC ports aren’t exposed in sidecar mode. That’s why your FastAPI app can’t connect over gRPC.

The good news: you don’t actually need gRPC here.

What still works:

  • Weaviate’s REST API on port 8080 remains fully available in sidecar containers.

  • All standard operations (inserts, queries, management) work fine over REST.

What to do:

  • Point your FastAPI client at http://localhost:8080.

  • In Weaviate’s config, set ENABLE_GRPC=false to skip gRPC entirely (avoids startup warnings).

  • Both containers in a sidecar share the same network namespace, so localhost:8080 will be reachable.

Quick check:
Inside your FastAPI container, run:

curl http://localhost:8080/v1/.well-known/ready

If you see a healthy response, your app can talk to Weaviate over REST just fine.

Azure did recently remove the YAML-based multi-container option, so sidecar + REST API is the intended path forward.

3

now on azure getting as

2025-09-10T08:46:08.2959156Z {“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“level”:“warning”,“log_level_env”:“”,“msg”:“log level not recognized, defaulting to info”,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:08.2960768Z {“action”:“startup”,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“level”:“info”,“msg”:“Feature flag LD integration disabled: could not locate WEAVIATE_LD_API_KEY env variable”,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:08.2960888Z {“action”:“startup”,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“default_vectorizer_module”:“text2vec-openai”,“level”:“info”,“msg”:“the default vectorizer modules is set to \“text2vec-openai\”, as a result all new schema classes without an explicit vectorizer setting, will use this vectorizer”,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:08.3059019Z {“action”:“startup”,“auto_schema_enabled”:true,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“level”:“info”,“msg”:“auto schema enabled setting is set to \“true\””,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:08.3469539Z {“action”:“memberlist_init”,“bind_port”:7946,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“error”:“Failed to get final advertise address: No private IP address found, and explicit IP not provided”,“hostname”:“node1”,“level”:“error”,“msg”:“memberlist not created”,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:08.347004Z {“action”:“startup”,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“error”:“create member list: Failed to get final advertise address: No private IP address found, and explicit IP not provided”,“level”:“error”,“msg”:“could not init cluster state”,“time”:“2025-09-10T08:46:08Z”}

2025-09-10T08:46:16.6956404Z 2025-09-10 08:46:16,652 app.services ERROR: Failed to connect to Weaviate. Connection to Weaviate failed. Details: Error: [Errno 99] Cannot assign requested address.

2025-09-10T08:46:16.6958252Z Is Weaviate running and reachable at http://localhost:8080?

for local i was using below yml file

using openai

services:

weaviate:

command:

- --host

- 0.0.0.0

- --port

- '8080'

- --scheme

- http

image: cr.weaviate.io/semitechnologies/weaviate:1.30.0

ports:

- 8080:8080

- 50051:50051

volumes:

- weaviate_data:/var/lib/weaviate

restart: on-failure:0

environment:

  QUERY_DEFAULTS_LIMIT: 20000

  AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'

  PERSISTENCE_DATA_PATH: '/var/lib/weaviate'

  DEFAULT_VECTORIZER_MODULE: 'text2vec-openai'

  ENABLE_MODULES: 'text2vec-openai'

  CLUSTER_HOSTNAME: 'node1'

  ENABLE_GRPC: 'false'

  OPENAI_APIKEY: 'sk-proj-0DGPLWqIXNKfVAmt1WB2bu76ZXdTj_ooT3BlbkFJQQ3Uvq1YXyRsRAyD2ZvDy1zIosjD1IUOIzQFnA7g65cridFnRdx3_Guem5A0Nx1PptZewAiwcA'

volumes:

weaviate_data:

i have added these as env variable in sidecar container, what might be the issue

4

hi @Abhishek_Joshi !

I couldn’t find anything related to ENABLE_GRPC. :thinking:

Also, GRPC is fundamental to a lot of operations.

Regarding the error, Weaviate was not able to be up and running:

2025-09-10T08:46:08.3469539Z {“action”:“memberlist_init”,“bind_port”:7946,“build_git_commit”:“b7b7715”,“build_go_version”:“go1.22.12”,“build_image_tag”:“v1.30.0”,“build_wv_version”:“1.30.0”,“error”:“Failed to get final advertise address: No private IP address found, and explicit IP not provided”,“hostname”:“node1”,“level”:“error”,“msg”:“memberlist not created”,“time”:“2025-09-10T08:46:08Z”}

It seems it was not able to identify itself with a private IP. So your app was not able to connect to it.

Running Weaviate as a sidecar is not something we have explored, to be honest.

1 Like
5

Thanks for the clarification @DudaNogueira You’re absolutely right I was thinking of a different containerized service that had a gRPC toggle.

Looking at the logs, the core issue is the memberlist clustering initialization failure in the Azure sidecar environment:

"Failed to get final advertise address: No private IP address found, and explicit IP not provided"

Since this is a single-node deployment in a sidecar, the fix is to avoid clustering issues and make sure Weaviate binds cleanly. For Azure sidecars, you can try:

environment:
QUERY_DEFAULTS_LIMIT: 20000
AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
DEFAULT_VECTORIZER_MODULE: 'text2vec-openai'
ENABLE_MODULES: 'text2vec-openai' # Single-node clustering config CLUSTER_HOSTNAME: 'node1'
OPENAI_APIKEY: 'your-api-key'

With this configuration, Weaviate will run in single-node mode without trying to initialize clustering. Since both containers share the same localhost network namespace in Azure sidecars, your FastAPI service should be able to connect via:

http://localhost:8080

The clustering error is what’s stopping Weaviate from starting fully, which is why your app couldn’t reach it.