Add healthcheck for database container

[Phu2]

With this change strapi will only start up if database is healthy/ready.
If the database is not ready starting up using docker compose up will end with

✔ strapi-rpb Built 0.0s
✔ Network strapi-rpb Created 0.1s
✘ Container strapi-rpbDB Error 1.1s
✔ Container strapi-rpb Created 0.1s

whereas a successful start up will be confirmed by

✔ strapi-rpb Built 0.0s
✔ Network strapi-rpb Created 0.2s
✔ Container strapi-rpbDB Healthy 5.9s
✔ Container strapi-rpb Started 6.0s

Why?
As just happened the other day, creating a file data/postgresql.conf.orig on the host machine as root will make strapi start with errors like:

strapi-rpbDB | chown: /var/lib/postgresql/data/postgresql.conf.orig: Operation not permitted strapi-rpb | [2025-02-04 07:03:20.350] debug: ⛔️ Server wasn't able to start properly. strapi-rpb | [2025-02-04 07:03:20.352] error: getaddrinfo EAI_AGAIN strapi-rpbDB strapi-rpb | Error: getaddrinfo EAI_AGAIN strapi-rpbDB strapi-rpb | at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:dns:109:26)

It is not immediately obvious what went wrong because the the docker compose up command doesn't show any errors.

[fsteeg]

We should merge the current main branch here, apply these changes to the new docker-compose-test.yml, and deploy to test.

[Phu2]

Deployed to test, docker compose -f docker-compose-test.yml ps shows:

NAME                   IMAGE                  COMMAND                  SERVICE        CREATED          STATUS                    PORTS
strapi-rpb             strapi-rpb:latest      "docker-entrypoint.s…"   strapi-rpb     18 seconds ago   Up 13 seconds             0.0.0.0:1339->1337/tcp
strapi-rpb-adminer-1   adminer                "entrypoint.sh php -…"   adminer        18 seconds ago   Up 13 seconds             0.0.0.0:8080->8080/tcp
strapi-rpbDB           postgres:12.0-alpine   "docker-entrypoint.s…"   strapi-rpbDB   18 seconds ago   Up 15 seconds (healthy)   0.0.0.0:5434->5432/tcp

See the healthy status of the database container.

[fsteeg]

I don't understand why the changes in the three files are so different: no condition: service_healthy in docker-compose-test, no healthcheck in docker-compose-prod, both condition: service_healthy and healthcheck in docker-compose.

[Phu2]

You are spot on. Somehow i've messed up the files when editing in compare mode of my editor. Re-deployed to test.

[fsteeg]

Looks good, will deploy on prod with next deployment there.

(One note on our test deployment process: I prefer to merge the feature branch into main on test, instead of switching the branch on test. That way, we can deploy multiple feature branches at the same time, without losing changes from another branch.)

[Phu2]

(One note on our test deployment process: I prefer to merge the feature branch into main on test, instead of switching the branch on test. That way, we can deploy multiple feature branches at the same time, without losing changes from another branch.)

Alright, i will merge into main from now on.