Metrics
Metrics show CPU, memory, network and restarts over time, for your whole organisation, one deploy, one pod, or one dedicated node.
Where to find metrics#
| Scope | Open from | Metrics |
|---|---|---|
| Organisation | Observability→Metrics | CPU, Memory, Net In, Net Out, one total line per metric across all of your workloads |
| Deploy | Deployments→⋮→View metrics | CPU, Memory, Net In, Net Out, Restarts |
| Pod | Infrastructure→Pods→⋮→View metrics | CPU, Memory, Net In, Net Out, Restarts |
| Dedicated node | Infrastructure→Nodes→⋮→View metrics | CPU, Memory, Net In, Net Out for the whole machine |
The organisation-wide Metrics page shows one card per metric, labelled Org total · last <range>, with ranges of 1h, 6h, 24h (the default) and 7d. The Overview page also shows a CPU usage chart for the organisation. It reads No usage yet — create a deploy to see CPU here. until something is running.
Using the metrics viewer#
The deploy, pod and node views open as a drawer. Pick a metric, then a time range: 5m, 15m, 1h (the default), 6h or 24h. The chart refreshes on its own every 20 seconds while the drawer is open.
- CPU
- CPU used, in cores.
- Memory
- Memory in use.
- Net In / Net Out
- Network traffic received and sent.
- Restarts
- How many times containers have restarted. Pods and deploys only. A node has no restarts.
Usage against allocation#
For a deploy or pod, the CPU and Memory charts also show the Current allocation, the build's request. When usage runs close to it, Increase CPU or Increase RAM opens the Adjust Resources drawer with that resource already one step higher, ready to confirm. See CPU and memory.
Restarts#
A deploy whose pods keep restarting shows a climbing Restarts count. The dashboard flags three or more restarts. It doesn't show why a container restarted, so check logs for the last lines before each restart, and memory usage against the allocation.
Node metrics are dedicated-only#
Node metrics are totals for the whole machine. A shared node runs other organisations' workloads too, so its totals aren't available and the API responds 403. On shared compute, read the deploy or pod metrics instead.
Metrics over the API#
GET /v1/metric-types
GET /v1/deploys/{deploy_id}/metrics?metric=cpu&range=1h&step=&pod=
GET /v1/pods/{pod}/metrics?metric=memory&range=6h
GET /v1/metric-types/nodes
GET /v1/nodes/{node_id}/metrics?metric=cpu&range=24h
GET /v1/organizations/{organization_id}/metrics?metric=cpu&range=1hResponses are { "metric", "series": [{ "labels": { "pod", … }, "points": [{ "timestamp", "value" }] }] }.