Skip to navigation

Monitoring

The Monitoring tab is a set of charts on each VM’s page that shows how busy the VM is: CPU, memory, disk space, disk activity and network traffic, for up to the last 7 days. Use it to see whether a VM is too small or too big, why it’s slow, and how much outbound data it has used this month.

When to use it

You want to know…Look at
Whether the VM needs more vCPUs or memoryMonitoring (this page)
How much outbound data the VM used this monthMonitoring → Network Traffic → Monthly outbound
What was done to the VM, by whom, and whether it worked (create, power actions, restores)Activity logs
GPU usage, GPU memory, temperature and powerThe GPU Metrics tab of a GPU VM
What’s happening inside a programTools inside the VM, such as top, htop or your app’s logs

How it works

  1. IBEE measures CPU, memory, disk activity and network traffic from outside the VM, on the host it runs on. This needs nothing installed.
  2. For Storage Utilization (how full the disk is), IBEE asks the QEMU guest agent, a small helper program inside the VM. IBEE’s Linux images install it at first boot.
  3. Samples are stored and shown as charts. The page refreshes itself every 30 seconds.

Before you begin

  • The VM must exist and, for live values, be Running. Charts keep their history while the VM is stopped.
  • For Storage Utilization, the guest agent must be running inside the VM. Check with systemctl status qemu-guest-agent (Linux).
  • Any workspace member who can see the VM can see its charts.

Open the charts

1

Open the VM

In the portal sidebar, open Infrastructure → Cloud VMs and click the VM.

2

Open Monitoring

Click the Monitoring tab. The header shows when the latest sample was taken and the interval between points.

3

Pick a time range

Click 30m, 1h, 6h, 24h or 7d. Click Refresh to load the newest values straight away.

Time ranges

RangeOne point every
30m1 minute
1h1 minute
6h5 minutes
24h15 minutes
7d2 hours

Longer ranges average more samples into each point, so short spikes look smaller. Use 30m or 1h to see them.

The charts

ChartWhat it showsNeeds the guest agent?
CPU UtilizationHow busy the VM’s vCPUs are, as a percentage of all of themNo
Memory UtilizationMemory in use compared with the plan’s memoryNo
Storage UtilizationHow full the VM’s filesystems are, as a percentage and as used / totalYes. Without it the chart says “Requires QEMU guest agent inside the VM”.
Disk ThroughputData read from and written to disk per secondNo
Disk IOPSDisk reads and writes per second (IOPS: input/output operations per second)No
Network TrafficInbound and outbound data per second, plus Monthly outbound for the current monthNo

Reading the charts

  • CPU above 80% for hours at a time: the VM is short of processor power. Check the 24h range before you resize.
  • Memory close to 100% and the VM is slow: it’s short of memory. Move to a bigger plan, or a Standard plan, which has 4 GiB per vCPU.
  • Storage above 90%: clean up, or attach a volume. A full root disk can stop the VM’s services.
  • High Disk IOPS while CPU is low: the app is waiting on the disk.
  • Monthly outbound close to your plan’s allowance: data over the allowance is charged per GB. See Instance types for each plan’s allowance.

Get the same data from the API

GPU VMs use the same paths under /compute/gpu-vms/{vm_id}.

curl "https://api.ibee.ai/v1/compute/cloud-vms/VM_ID/metrics?workspace_id=$IBEE_WORKSPACE_ID" \
-H "Authorization: Bearer $IBEE_TOKEN"
RequestReturns
GET /compute/cloud-vms/{vm_id}/metricsThe latest values
GET /compute/cloud-vms/{vm_id}/metrics/timeseriesThe chart data for a range
GET /compute/cloud-vms/{vm_id}/metrics/bandwidthMonthly network usage
GET /compute/cloud-vms/{vm_id}/eventsThe VM’s recent events (API only)

With the CLI: ibee vms metrics VM_ID, ibee vms metrics-timeseries VM_ID, ibee vms bandwidth VM_ID and ibee vms events VM_ID.

Limits

Value
Longest range in the portal7 days
Finest interval1 minute
Page refreshEvery 30 seconds

Troubleshooting

You seeWhat it meansWhat to do
Storage Utilization says “Requires QEMU guest agent inside the VM”The guest agent isn’t installed or isn’t runningOn Ubuntu or Debian: sudo apt-get install -y qemu-guest-agent && sudo systemctl enable --now qemu-guest-agent. On Rocky, AlmaLinux or Fedora, use sudo dnf install -y qemu-guest-agent.
”No disk throughput data yet” or “No disk IOPS data yet”The VM is new or has been stoppedWait for a few samples while the VM is Running.
”Historical charts are available but the latest sample has not refreshed recently.”New samples have stopped arriving, for example because the VM is stoppedStart the VM, or click Refresh. If it persists while the VM is running, contact support.
A spike you saw in top doesn’t showThe range averages it awaySwitch to 30m.