Save any received jobs in a postgresql database

[lfse-slafleur]

• Is necessary to link OMOTES job id with Celery task id so the Celery result can be retrieved once a job is done.
• Is necessary to show the current state.
• Delete them once the job is finished and the result is forwarded to the OMOTES sdk.

Please take care of ordering: Any handler in the orchestrator that processes an AMQP message will acknowledge that message the moment the function returns. The handler should only succeed and acknowledge the message once any postgresql transactions have completed. If this invariant is broken, we will lose work/state in edge-case situations.

[lfse-slafleur]

Unsure if necessary at the moment. Currently all work for orchestrator is kept in durable queues and no info needs to be saved. Perhaps we do not need postgresql for this purpose.

[lfse-slafleur]

Necessary for cancel and timeout functionalities.