[CoR] Default Cosmos DB config fails

Nobody has claimed this yet.

With the default setup, the cosmosdb container never starts up and the compose command eventually times out. Copilot is able to fix it and explains the issue:

AZURE_COSMOS_EMULATOR_IP_ADDRESS_OVERRIDE=127.0.0.1 — this forces the emulator's internal certificate to bind to loopback, which can cause conflicts inside the container's network namespace. Removing it lets the emulator use its default address.

The emulator is hitting a fatal error (0xC0000001) in its internal AppLoader. This is a known issue with the Cosmos DB Linux emulator when using bind-mounted volumes on Docker Desktop for Windows — the emulator's internal SQL PAL layer has compatibility issues with NTFS-backed bind mounts.

The fix is to replace the bind mount with a named Docker volume (or remove data persistence entirely since the migration service recreates containers on startup anyway)

It makes changes to the compose config

full docker-compose.yml

services:
  azurite:
    image: mcr.microsoft.com/azure-storage/azurite
    command: azurite --blobHost 0.0.0.0 --queueHost 0.0.0.0 --tableHost 0.0.0.0 --skipApiVersionCheck
    ports:
      - "10000:10000"
      - "10001:10001"
      - "10002:10002"
    volumes:
      - ./.azurite:/data
    restart: unless-stopped

  cosmos-emulator:
    image: mcr.microsoft.com/cosmosdb/linux/azure-cosmos-emulator:latest
    ports:
      - "8081:8081"
      - "10250:10250"
      - "10251:10251"
      - "10252:10252"
      - "10253:10253"
      - "10254:10254"
    environment:
      - AZURE_COSMOS_EMULATOR_PARTITION_COUNT=3
    volumes:
      - cosmos-data:/tmp/cosmos/appdata
    healthcheck:
      test: [ "CMD-SHELL", "curl -ks https://localhost:8081/_explorer/emulator.pem || exit 1" ]
      interval: 10s
      timeout: 5s
      retries: 30
      start_period: 60s
    restart: unless-stopped

  db-migrate:
    image: node:24-slim
    working_dir: /app/services/microblog-api
    depends_on:
      cosmos-emulator:
        condition: service_healthy
    volumes:
      - ./:/app:ro
      - ./node_modules:/app/node_modules:ro
    environment:
      COSMOSDB_CONNECTION_STRING: "AccountEndpoint=https://cosmos-emulator:8081/;AccountKey=C2y6yDjf5/R+ob0N8A7Cgv30VRDJIWEHLM+4QDU5DE2nQ9nDuVTqobD4b8mGGyPMbIZnqyMsEcaGQy67XIw/Jw=="
      COSMOSDB_DATABASE_NAME: "microblog"
      NODE_TLS_REJECT_UNAUTHORIZED: "0"
    entrypoint: [ "npm", "run", "migrate" ]
    restart: "no"

volumes:
  cosmos-data:

Start by locating the default docker-compose.yml used by the project and inspect the cosmos-emulator service configuration, especially its environment and volume settings. Run the compose command on the default setup and verify that the Cosmos DB emulator becomes healthy and db-migrate completes without timing out.

Written by the indexing model from the issue text.

Description

Contributor guide

First steps

Research direction

Assessment