Monitoring virtual machines¶

Basic usage¶

Basic monitoring of the VMs in your project is available via our Grafana pages. You can see your project at the "magic URL" job.c3se.chalmers.se/cirrus/PROJECT_UUID and an individual VM at job.c3se.chalmers.se/cirrus/VM_UUID. If you have set up the OpenStack CLI you can see your project and server UUIDs using

$ openstack project list -fvalue -c{id,name}
f05816d31b33... C3SE X15/YY-Z

The first, alphanumeric string is the project UUID, the second is the project name.

So https://job.c3se.chalmers.se/cirrus/f05816d31b33... would redirect to your Grafana page.

$ openstack server list -fvalue -c{id,name}
21fd882b-93... first-vm
960eb241-e90... second-vm

The alphanumeric string with dashes is the VM UUID, the second is the name you gave when you created it. So, https://job.c3se.chalmers.se/cirrus/21fd882b-93... would redirect to the Grafana page of first-vm, and https://job.c3se.chalmers.se/cirrus/960eb241-e90... would redirect to the Grafana page of second-vm.

These pages will show a basic view of the CPU usage and memory allocation of your machines. The memory allocation is not necessarily accurate as it includes cached memory.

Setting up your own monitoring¶

You can deploy a monitoring software like zabbix to monitor your VM:s. You deploy it and collect information using the project-internal network 10.10.10.0/24 that OpenStack automatically sets up for you. Carefully consider whether to give such monitoring software a public IP at all. External access can be useful for troubleshooting if your point-of-access VM is not working properly. However, monitoring software can typically execute code remotely, and it is therefore very important to properly firewall and secure it if publicly exposed.

If you do give it a public IP, make sure the firewall rules are as restrictive as possible.