Kibana canvas timing out after upgrade

Elastic Stack Kibana
1

Hi,
I've just upgraded from elasticsearch+kibana 8.13.3 to 8.16.1.
The upgrade seemed to go smoothly, but now I notice my canvas dashboards show clocks for ages and many elements (using essql with elasticsearch) then fail with request timeout.
Even the flight overview using demo data fails to render lots of elements due to timeouts.
Looking at stack monitoring, nothing jumps out for either kibana or elasticsearch (everything marked as healthy) and I don't see anything obvious in kibana logs either.
Tried doing some queries against all the elasticsearch nodes and they all came back nice and quick. Maybe worth mentioning this is in an airgapped environment and everything was working fine immediately prior to upgrade.
It's running in vmware and I shut everything down and took snapshots prior to starting the upgrade so I am able to roll the whole of the environment back if needs be to try again.

Thanks
Steve

2

Just for closure:
We have 3 elasticsearch nodes. In kibana.yml they're specified in elasticsearch.hosts. The last one, the IP was a typo. Bizarre that it's never been an issue before, maybe the round-robin logic for which hosts it uses has been updated?
What was strange was the same places would fail (e.g. on a canvas, 2 out of 3 elements would work, the third would fail, also retrieving api keys would always fail).
Now updated and all working well.
What would be useful for future troubleshooting is:
Does anyone know how to get kibana to include the host and port that it's trying to connect to in the log file?
Tried setting things to trace level but I couldn't work out how to get it to include the host and port, only the request itself?

Cheers

Steve