Questions on Redis and Pipelines in Qovery

1-Qovery supports a single AWS Elasticache REDIS node with daily backups (no Multi-AZ supprt). Qovery does not support MemCached instances either, so this means that our in-memory Redis data is not persisted in a backing storage and we would lose our data if the Redis instance were to be restarted. It appears that AWS MemCached single node provides persisting storage capabilities at a cost comparable to the single node in-memory only REDIS. Is there a reason why Qovery does not support MemCached? Are my statements above accurate? Any ideas how we can support in-memory data persistence in a cost-effective way?
2-Looking into pipelines to speed up deployments. I see that options available for application deployments is either to point a registry and then be able to specify startup commands OR choose build from GIT and lose the ability to specify startup commands. That basically means we have the exact same docker instructions duplicated with different startup commands for each the applications.
ideally, would like to
-Do a single build from GIT
-Verify build is successful
-Run Rails Database Migrations
-Use same image built and parallelize the deployment of all applications, while specifying different CMD commands for each application
CMD [“bundle”, “exec”, “puma”, “-C”, “./config/puma.rb”]
CMD [“bin/rails”, “resque:scheduler”]
CMD [“bin/rails”, “resque:work”]
CMD [“bundle”, “exec”, “clockwork”, “lib/clock.rb”]
Is there a way to set this up via API or terraform ?