> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vms.verolabs.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Probe: k8s_mon

> Monitor Kubernetes workloads, pods, nodes, and services

<Warning>
  This probe is on the development roadmap.
</Warning>

## Description

The `k8s_mon` probe monitors Kubernetes resources: Deployment, DaemonSet, StatefulSet (replicas), Pod (readiness, restarts), Node (readiness), and Service (existence). Optional CPU/memory via metrics-server.

Build tag `vms_heavy`. Uses in-cluster config + ServiceAccount. Requires RBAC get/list on the target resource.

## Config fields

| Field           | Type     | Required | Default     | Description                                                                                   |
| --------------- | -------- | -------- | ----------- | --------------------------------------------------------------------------------------------- |
| `type`          | `string` | Yes      | —           | Must be `"k8s_mon"`                                                                           |
| `namespace`     | `string` | No       | `"default"` | Kubernetes namespace                                                                          |
| `workloadKind`  | `string` | Yes      | —           | Resource kind: `"Node"`, `"Pod"`, `"Deployment"`, `"DaemonSet"`, `"StatefulSet"`, `"Service"` |
| `workloadName`  | `string` | \*       | —           | Specific resource name. One of `workloadName` or `labelSelector` required                     |
| `labelSelector` | `string` | \*       | —           | Label selector (e.g. `"app=core-api"`)                                                        |
| `collectStats`  | `bool`   | No       | `false`     | Collect CPU/memory per pod via metrics-server                                                 |

## Metrics

### Deployment / DaemonSet / StatefulSet

| Metric                           | Type   | Labels                          |
| -------------------------------- | ------ | ------------------------------- |
| `vms.k8s_mon.status`             | status | `kind`, `namespace`, `workload` |
| `vms.k8s_mon.workload.available` | gauge  | `kind`, `namespace`, `workload` |
| `vms.k8s_mon.workload.desired`   | gauge  | `kind`, `namespace`, `workload` |

### Pod

| Metric                      | Type    | Labels                     |
| --------------------------- | ------- | -------------------------- |
| `vms.k8s_mon.pod.ready`     | gauge   | `kind`, `namespace`, `pod` |
| `vms.k8s_mon.restart_count` | counter | `kind`, `namespace`, `pod` |

### Stats (collectStats = true)

| Metric                       | Type  | Unit       | Labels                     |
| ---------------------------- | ----- | ---------- | -------------------------- |
| `vms.k8s_mon.cpu_millicores` | gauge | millicores | `kind`, `namespace`, `pod` |
| `vms.k8s_mon.memory_bytes`   | gauge | bytes      | `kind`, `namespace`, `pod` |

## Example config

```json theme={null}
{
  "type": "k8s_mon",
  "namespace": "core",
  "workloadKind": "Deployment",
  "workloadName": "core-api",
  "collectStats": true
}
```

## Notes

* Max **100 pods/nodes** emitted per run.
* Missing RBAC → `unknown` status with `vms.agent.permission_missing{capability=kubernetes_rbac}`.
* `collectStats` requires metrics-server installed. If absent, stats are silently skipped.
* Build tag `vms_heavy` — not in default agent package.