Hi,

Im trying to achieve zero-downtime deployment in my Laravel app and after months of trying different approaches, Im able to get the 502 downtime down to second, but still have not got it working (zero downtime).

For the context, Im using https://github.com/jackbrycesmith/laravel-caprover-template and I have moved everything from entrypoint.sh to Dockerfile (followed this discussion). There are a few stuff remain in the entrypoint.sh that cant be moved (here is my current file):

#!/bin/sh

php artisan config:cache
php artisan route:cache
php artisan event:cache

php artisan migrate --no-interaction --force

supervisord -c $LARAVEL_PATH/.deploy/production/config/supervisor.conf

The problem is: if the new version/deployment does not have migration to run, the downtime is almost zero (which is great). Otherwise, the downtime will equal to the time it takes to run the migration.

I end up with putting RUN sleep 10 into my Dockerfile as a leeway but that is not a great solution.

So my next attempt is to write a script and put that at the end of Dockerfile (or after running entrypoint.sh) to manually and constantly check the current being deployed container is up and running:

if [ "$(docker inspect --format='{{json .State.Health.Status}}' <container_name_or_id>)" = '"healthy"' ]; then
    echo "Container is healthy"
else
    echo "Container is unhealthy"
fi

The question is: how to get <container_name_or_id> var?

I asked ChatGPT and got this answer but not sure if there is better approach?

If you really need to access the container ID within the Dockerfile (which is not typical), you could technically write a script to extract the container ID at runtime and write it to a file during container startup. However, this approach is not recommended as it goes against the typical Docker container workflow and introduces unnecessary complexity.

P/S: or is there a way to fire local request to the app URL to check if its 200 OK?

Cheers,