[Breaking] drop support for unclustered setup#611
Conversation
|
Goodbye, unclustered 🫡 |
moskyb
left a comment
There was a problem hiding this comment.
oh whoops i commented without approving!
b92ee85 to
ec6ab64
Compare
sj26
left a comment
There was a problem hiding this comment.
Nice! Not sure if you were waiting for me - but please go ahead!
DrJosh9000
left a comment
There was a problem hiding this comment.
I think we have consensus.
|
As a customer: Why? This doesn't have any advantage for us, so now we need to migrate all our agents/pipelines/etc over to this new clustered agent config, for zero benefit. Slipping this in just makes us unable to upgrade. |
|
Hi @willhughes-au. Sorry, we understand this is frustrating. We can't continue offering an unclustered mode. The implementation is completely different, and stops us shipping new features. We had to make the call that future versions of this agent stack for Kubernetes, which will need to leverage the clustered implementation, rely on clusters. For customers like yourself happily using unclustered mode, we want to make it a transparent transition to move to clusters, and if you're operating in one cluster then it's basically invisible. Everything should be the same, there's just a cluster wrapping the outside in the implementation. We don't have an easy button for this transition yet. But reach out and we'll make it happen for you: [email protected], CC me: [email protected] Same offer applies else who might stumble on this 🙏 |
THIS IS A BREAKING CHANGE
Linear card: PIPE-1114
After this change, unclustered agent token will be rejected, causing controller to abort.
This PR is stacked on top of: