I have a custom Oracle Database docker image that I need to use for my builds, but in order for the container to start properly, I need to set the following flags on the docker run command:

#!bash

--shm-size 1g
--ulimit nofile=1024:65536
--ulimit nproc=16384
--ulimit stack=10485760:33554432
--ulimit memlock=137438953472

I can get away with just setting --shm-size, but being able to set the ulimit params would allow me to commit and use the image rather than needing a separate process to setup the production image.

I've tried setting up my container as a service in my pipeline, but it appears not to be running when my build gets to the part where it relies on the database. As a side request, you should also provide deeper logging for the docker environment used by builds so we can see what is actually happening. I can only guess (with 99.9% certainty based on experience outside Bitbucket Pipelines) what is happening without it.