Thank you, Joao, This feature would be of a great help for me for sure. Can't wait for the 4.19.1 to be released!
In the mean time, i will check out the cloud-usage database and purge the old records. Perhaps that would help too. Cheers ----- Original Message ----- > From: "João Jandre Paraquetti" <j...@scclouds.com.br> > To: "users" <users@cloudstack.apache.org> > Sent: Tuesday, 18 June, 2024 19:09:30 > Subject: Re: Disabling VM metrics > Hello, Andrei > > Currently you cannot disable listing metrics when using the UI, this > behavior has been discussed and changed with PR > https://github.com/apache/cloudstack/pull/8782. On the next minor > release (4.19.1) there will be a configuration to let you change the > behavior of the `listVirtualMachines` API so that it does not return the > metrics by default. > > Furthermore, if you do not care for the metrics, you could change the > `vm.stats.max.retention.time` value to 1, so that the VM's metrics are > only kept for one minute; thus minimizing the amount of metrics being > listed. Beware that setting this value to 0 or lower will disable the > metrics cleanup, so 1 is the lowest value you can go. > > Also, if you have too many metrics collected, you might run into the > issue that is described here > https://github.com/apache/cloudstack/pull/8740, where due to the amount > of metrics that ACS tries to delete in a single query, the query always > times out and this snowballs into a huge amount of metrics on your DB. > The linked PR solves this adding another configuration to limit the > amount of metrics deleted per query, hopefully it will be in by > 4.19.1.0. Until then, if you notice that the metrics are not being > deleted, you might have to manually delete the old ones on the DB. > > Best regards, > > João Jandre > > On 6/18/24 03:40, Andrei Mikhailovsky wrote: >> Hello everyone, >> >> Could someone recommend me a way to disable the collection of vm metrics? It >> is >> currently taking a long time (10-20 seconds) for the vm information to show >> up >> when clicking on a vm. Also, when trying to attach a volume(s) to a vm, the >> full list of vms is not populating (my guess due to the timeout when >> retrieving >> vm metrics). I've tried reducing the length of collected data in the >> cloudstack >> settings, as suggested by someone on this list, but this did not solve the >> problem. >> >> Currently, the vm metrics data are not being used and their collection is >> causing too much inconvenience and at times makes the gui unusable. Could >> someone suggest the way the collection of vm metrics data could be switched >> off? >> >> Many thanks >> >> Andrei
