Loïc Mathieu

@dliubarskyi Yes, #3570 would be enought for us to provides a standard authentication mechanism. I subscribe to it and hope it would be integrated soon.

Yes, we need investiguation as we may not have the correct execution context. As a side note, we can relax the requirement to specify the namespace of the subflow and...

We restict flowId, triggerId, namespace, ... and a bunch of other fields to 250 chars in the JDBC runner (not in the Kafka one) but didn't add any size limit....

This has alread been discussed on the mailing list: https://groups.google.com/d/msg/quarkus-dev/01GSKgzQ9dA/GjgE2c7nAwAJ

@alesj it needs to be pushed to the `EntityManager` as it's an hibernate feature. I have a working prototype using a CDI interceptor

@stefanorg the PR is for readonly hibernate session (called readonly transaction by some other framework), it's a different thing as readonly connection to route to a different tenant. You should...

@jmfinelli it has nothing to do with Narayana itself but as it'll need coordination between the transaction manager and the entity manager it will impact the Narayana extension. The transactional...

@jmfinelli no, the proposed implementation was refused but we keep the issue open for feedback and may propose something at some time.

@jmfinelli I removed the lable.

It's not stored, why the issue.