I will definitely have to downgrade.
Camunda 8.9+ has many problems, I don’t know if it’s just a matter of using self-managed with Docker on a Linux server, but many things are inconsistent.
Deploying process definitions shows success, but sometimes you have to try several times for them to be reflected in the processes published in …:8080/tasklist/processes?continue
Logouts in Operate and Tasklist don’t work.
Even downloading connectors, if you don’t refresh the page, it often doesn’t work.
Starting an instance with deploy & run is another issue; it often succeeds but doesn’t appear in Operate.
Etc Etc Etc
Thank you for sharing your experience. Feedback like this is genuinely valuable to us.
I’m from the QA Engineering team at Camunda, and I want to reassure you that 8.9 is actively used and tested by many teams, including ours. The issues you describe
process definitions not reflecting in Tasklist
logout behavior
connectors requiring a page refresh
instances not appearing in Operate
are not things we see consistently in our test environments, which makes me think there may be something specific to your setup that’s contributing to these symptoms.
That said, I don’t want to dismiss your experience. To help us understand whether this is a reproducible bug or a configuration/environment issue, it would be really useful if you could share detailed information about environment & configuration, component versions, log files and steps to reproduce.
For the logout behavior specifically - this is a known area we’ve been improving. In 8.9, RP-initiated logout (full IdP logout) was introduced but requires explicit configuration. If you’re using OIDC, it’s worth checking the logout configuration docs.
We’d love to help get to the bottom of this rather than see you downgrade. If you can share the details above, we can take a proper look.
Hints: Use the Ask AI feature in 