You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During some debugging with CSI and CAPV, we've figured out some sessions being opened a lot to vCenter and VAPI, and in fact we got some improvements from it.
One thing that was really helpful on CSI side was the fact that they have implemented the clusterID as part of the userAgent, so listing vc sessions (govc session.ls) allowed us to summarize how many connections each cluster/CSI was opening, if the session was stuck, if it was same userAgent and different nodes (some leader election issue?), etc.
I would like to request that we consider the same for vSphere CPI. I couldn't find on the code any reference on setting the clusterID as a config, so I wanted to understand if:
We can add this extra configuration on yaml config (and make it clear that right now this information is only used for userAgent)
We can mutate userAgent to, in case clusterID is set, being part of its composition
Thanks
Why is this needed?
Facilitate investigation on VC session leaks, stale connections, add better observation on who/what cluster is consuming connections
The text was updated successfully, but these errors were encountered:
What would you like to be added?
Hey folks,
During some debugging with CSI and CAPV, we've figured out some sessions being opened a lot to vCenter and VAPI, and in fact we got some improvements from it.
One thing that was really helpful on CSI side was the fact that they have implemented the clusterID as part of the userAgent, so listing vc sessions (govc session.ls) allowed us to summarize how many connections each cluster/CSI was opening, if the session was stuck, if it was same userAgent and different nodes (some leader election issue?), etc.
I would like to request that we consider the same for vSphere CPI. I couldn't find on the code any reference on setting the clusterID as a config, so I wanted to understand if:
Thanks
Why is this needed?
Facilitate investigation on VC session leaks, stale connections, add better observation on who/what cluster is consuming connections
The text was updated successfully, but these errors were encountered: