Migrated vSphere - need to update Grafana to new stack

Our Grafana instance was set up by someone who has since moved on to other employment. One of the panels we use in multiple dashboards is monitoring the CPU usage of the ESXi hosts in our vSphere cluster. We’ve just completed migrating our VMs to a new cluster, and I’m trying to figure out how to update Grafana to point to the new cluster instead of the old.

The only place I can see a reference to vCenter in our configs is in /etc/telegraf/telegraf.d/vcenter.conf - I changed the hostname there to point to the new cluster, and rebooted everything. But I’m still getting data from the old vCenter cluster on the dashboard, which really makes no sense to me.

Since I didn’t have a hand in setting this up originally, I’m sort of trying to reverse-engineer what we’ve got set up. Hoping someone might be able to point me in the right direction. We’re on Grafana 7.0.0.

You need to be patience, because the telegraf need time to scrape. You need to check the data inside the influxdb using influxdb admin tools.
As my experience, the Vcenter data need little bit time to display into the Grafana, depends on the time shift of the dashboard.
If you want to make sure, also check the telegraf log to see the telegraf getting the data from vcenter cluster.
You already at the right direction to change the vmware target in the telegraf, just need deeper digging into the telegrfaf log and influxdb data.


Do you have any tips/hints on checking the data within InfluxDB? I’m a bit of a newbie there - I got logged into the DB and selected the DB, but wasn’t really sure what to do after that.

You can use this for UI explorer for influxdb

To make sure just go to the telegraf machine and try to check with tail -f /var/log/telegraf/telegraf.log
For better info, restart the telegraf service from another shell and you see the telegraf will produce the log.

Hm, getting this when I try to use Time Series Admin:
“Error: write EPROTO 140595240908472:error:100000f7:SSL routines:OPENSSL_internal:WRONG_VERSION_NUMBER:…/…/third_party/boringssl/src/ssl/tls_record.cc:242: at WriteWrap.onWriteComplete [as oncomplete] (internal/stream_base_commons.js:87:16)” No clue on that one, I’m afraid.

I did find some other seemingly random vsphere-influxdb-go.json file at /etc/ that I needed to update a bit, and then everything worked as expected. No idea why that was hanging out in /etc instead of being in /etc/telegraf or /etc/influxdb. Since it’s just hanging out in the root, I don’t even really know whose file it is (though the name would certainly suggest that it’s influxdb’s file). But, at any rate, I think it’s all good now.