Connection terminated due to connection timeout error

まず、報告されている TypeORM、pg、Azure Database for PostgreSQL、AKS の構成を確認し、その後、SKU Standard_B4ms と Standard_D4as_v5 の間で動作を比較します。接続タイムアウトのスタックトレースと、Azure Accelerated Networking における差異の可能性を調査します。再現可能な原因と、検証済みの構成または依存関係の修正を特定できれば完了です。

We are using NestJS with TypeORM v0.3.25 and pg v8.11.3

Our TypeOrm config is:

const baseOrmConfig: PostgresConnectionOptions & TypeOrmModuleOptions = {
  type: 'postgres',
  host: process.env.DATABASE_HOST ?? 'localhost',
  port: process.env.DATABASE_PORT !== undefined ? Number(process.env.DATABASE_PORT) : 5432,
  username: process.env.DATABASE_USERNAME ?? 'postgres',
  password: process.env.DATABASE_PASSWORD ?? 'password',
  synchronize: false,
  logging: ['error'],
  autoLoadEntities: true,
  ssl: process.env.DATABASE_HOST !== undefined,
  extra: {
    // setting this to 0 to disable auto-disconnect cspell: disable-next-line
    idleTimeoutMillis: 60_000,
    keepAlive: true,
    keepAliveInitialDelayMillis: 60_000,
    max: 10,
    connectionTimeoutMillis: 30_000, // number of milliseconds to wait for connection
  },
};

We host this service on a Azure Kubernetes Service cluster, and are making connections to an Azure Database for PostgreSQL Flexible Server.

Recently, we upgraded our AKS Node SKUs from Standard_B4ms to Standard_D4as_v5, and started seeing a lot of the foll. error:

Error: Connection terminated due to connection timeout 
  at Client.<anonymous> (/usr/src/app/node_modules/pg-pool/index.js:251:17) 
  at Client.patchedClientConnectCallback (/usr/src/app/node_modules/@opentelemetry/instrumentation-pg/build/src/utils.js:309:12) 
  at /usr/src/app/node_modules/@opentelemetry/context-async-hooks/build/src/AbstractAsyncHooksContextManager.js:50:55 
  at AsyncLocalStorage.run (node:internal/async_local_storage/async_hooks:91:14) 
  at SentryContextManager.with (/usr/src/app/node_modules/@opentelemetry/context-async-hooks/build/src/AsyncLocalStorageContextManager.js:33:40) 
  at SentryContextManager.with (/usr/src/app/node_modules/@sentry/opentelemetry/build/cjs/index.js:1430:24) 
  at Client.contextWrapper [as _connectionCallback] (/usr/src/app/node_modules/@opentelemetry/context-async-hooks/build/src/AbstractAsyncHooksContextManager.js:50:32) 
  at Connection.<anonymous> (/usr/src/app/node_modules/pg/lib/client.js:145:18) 
  at Object.onceWrapper (node:events:632:28) 
  at Connection.emit (node:events:518:28) {   [cause]: Error: Connection terminated unexpectedly 
  at Connection.<anonymous> (/usr/src/app/node_modules/pg/lib/client.js:132:73) 
  at Object.onceWrapper (node:events:632:28) 
  at Connection.emit (node:events:518:28) 
  at Socket.<anonymous> (/usr/src/app/node_modules/pg/lib/connection.js:63:12) 
  at Socket.emit (node:events:518:28) 
  at TCP.<anonymous> (node:net:346:12) 
at TCP.callbackTrampoline (node:internal/async_hooks:130:17) } 

We reverted to using the Standard_B4ms SKU and the errors stopped. However, we do need to upgrade our SKUs. We did not have this error on our setup before this.

Any ideas what could be causing this error, and how we could fix this? The only possible explanation we could have found was that it could be related to Azure Accelerated Networking, but have no proof.