Monitor a VM

The Monitoring tab of a TIR Virtual Machine shows how much CPU, memory, disk, network, and storage the VM is using, and on a GPU plan, how hard the GPUs are working. TIR collects the figures for you, so there is nothing to install. Use the tab to catch a full disk or a busy processor before it slows your work.

Before you begin

You need a VM in the Running state, because the figures come from the running VM. See Stop, start, and restart a VM.

Open the Monitoring tab

In Manage VMs, select the VM’s name, then select the Monitoring tab.

Disk Usage at the top says how full the disk is, measured when you open the tab. To look further back than the last five minutes, choose 1h, 1d, 7d, or 30d, for example 7d to see whether a slowdown repeats at the same time each day. The refresh button loads the latest figures.

Read the figures

Figure	What it tells you
CPU Utilization, RAM Utilization	Whether the plan has enough vCPUs and memory
Uptime	How long the VM has run since it last booted
Network Bandwidth, Network Packet Rate	How much traffic the VM sends and receives
Storage IOPS, Storage Throughput	How busy the disk is
GPU Utilization, GPU Memory Used	On a GPU plan, how hard the GPUs work and how full they are
GPU Temperature, GPU Power	On a GPU plan, the GPUs’ heat and power draw
GPU Memory Bandwidth, Tensor Throughput, PCIe Throughput	On a GPU plan, how fast data moves through the GPUs

Max GPU Temperature is included when the GPU reports a value.

When the disk is filling up, resize the disk. When CPU or memory stays near the top, move the VM to a larger plan.

Read a VM’s logs

A VM keeps its logs inside its operating system, and you read them there. Sign in over SSH, as in Connect to a VM over SSH, and use the tools of the image, such as journalctl.

To get a warning when a figure crosses a limit, see Set up alerts for a VM. For help with a message on the tab, see Troubleshooting TIR Virtual Machines.