Error with PostgreSQL 15

[AndreMiras]

Is there an existing issue for this?

• I have searched the existing issues

Kong version ($ kong version)

3.4

Current Behavior

Error starting Kong with PostgreSQL 15

2023/08/21 13:38:42 [error] 1#0: init_by_lua error: /usr/local/share/lua/5.1/kong/init.lua:590: [PostgreSQL error] failed to retrieve PostgreSQL server_version_num: host or service not provided, or not known
stack traceback:
        [C]: in function 'assert'
        /usr/local/share/lua/5.1/kong/init.lua:590: in function 'init'
        init_by_lua:3: in main chunk
nginx: [error] init_by_lua error: /usr/local/share/lua/5.1/kong/init.lua:590: [PostgreSQL error] failed to retrieve PostgreSQL server_version_num: host or service not provided, or not known
stack traceback:
        [C]: in function 'assert'
        /usr/local/share/lua/5.1/kong/init.lua:590: in function 'init'
        init_by_lua:3: in main chunk

Expected Behavior

Instead I'm expecting the service to start and output the following logs (like with PostgreSQL 14):

2023/08/21 13:38:15 [notice] 1#0: using the "epoll" event method
2023/08/21 13:38:15 [notice] 1#0: openresty/1.21.4.1
2023/08/21 13:38:15 [notice] 1#0: OS: Linux 6.2.10-arch1-1
2023/08/21 13:38:15 [notice] 1#0: getrlimit(RLIMIT_NOFILE): 1073741816:1073741816
2023/08/21 13:38:15 [notice] 1#0: start worker processes
2023/08/21 13:38:15 [notice] 1#0: start worker process 1255
2023/08/21 13:38:15 [notice] 1#0: start worker process 1256
2023/08/21 13:38:15 [notice] 1#0: start worker process 1257
2023/08/21 13:38:15 [notice] 1#0: start worker process 1258
2023/08/21 13:38:15 [notice] 1#0: start worker process 1259
2023/08/21 13:38:15 [notice] 1#0: start worker process 1260
2023/08/21 13:38:15 [notice] 1#0: start worker process 1261
2023/08/21 13:38:15 [notice] 1#0: start worker process 1262
2023/08/21 13:38:16 [notice] 1255#0: *1 [lua] warmup.lua:106: single_dao(): Preloading 'services' into the core_cache..., context: init_worker_by_lua*
2023/08/21 13:38:16 [notice] 1255#0: *1 [lua] warmup.lua:147: single_dao(): finished preloading 'services' into the core_cache (in 0ms), context: init_worker_by_lua*

Steps To Reproduce

Using this docker-compose.yml:

version: "3.8"

services:
  kong-migrations:
    image: kong:3.4
    command: kong migrations bootstrap
    environment:
      KONG_PROXY_ERROR_LOG: /dev/stderr
      KONG_PG_USER: kong
      KONG_PG_DATABASE: kong
      KONG_PG_PASSWORD: kong
      KONG_PG_HOST: db
    depends_on:
    - db

  db:
    image: postgres:15
    environment:
      POSTGRES_PASSWORD: kong
      POSTGRES_USER: kong

  kong:
    image: kong:3.4
    environment:
      KONG_PROXY_ERROR_LOG: /dev/stderr
      KONG_PG_USER: kong
      KONG_PG_DATABASE: kong
      KONG_PG_PASSWORD: kong
      KONG_PG_HOST: db
    depends_on:
    - kong-migrations

Then:

docker compose up

Anything else?

We had a relatively similar issue with PostgreSQL 14 in the past, see #8259

[AndreMiras]

I'm not entirely sure how it happened as I could consistently reproduce using that docker-composer.yml file, but as I was tracking it down, it disappeared and I no longer see the issue.
Sorry for the noise 🤦‍♂️