We run `/health` checks on the server container every 5 seconds, which
makes the container logs hard to read. `healthcheck` still runs
correctly without the need for logs.