/cluster-health

Health check for Nutanix clusters covering nodes, alerts, and recent failed tasks

Boost this command

wyre-technology@wyre-technology

GitHub

From plugin

msp-claude-plugins

46200 skills146 agents200 commands4 MCP

Install

$ npx -y skills add wyre-technology/msp-claude-plugins --agent claude-code

How it fires

How this command gets triggered: by you, by Claude, or both.

Fires itselfClaude auto-loads it when your prompt matches the work.
You can call itInvoke it directly when you want it.
Slash command/cluster-health

Context preview

What this command does when you run it.

Health check for Nutanix clusters covering nodes, alerts, and recent failed tasks

Command definition

cluster-health.md

description: Health check for Nutanix clusters covering nodes, alerts, and recent failed tasks
argument-hint: "[cluster_name]"
arguments: [cluster_name]

Cluster Health

At-a-glance health for one or all Nutanix clusters: node state, open alerts by severity, and recent failed Prism Central tasks.

Prerequisites

Nutanix connected in the Conduit gateway with valid Prism Central credentials

Steps

1. **List clusters**

Discover the list-clusters operation via `listOperations` with `namespace="clustermgmt"`, then call `clustermgmt_execute`. Scope to the named cluster if given; otherwise cover all.

2. **Per-cluster detail**

For each cluster, get it by `extId` and list its hosts via `clustermgmt_execute` — capture node count, version, and any degraded host state.

3. **Open alerts**

Discover the list-alerts operation in `monitoring` and call `monitoring_execute` filtered to unresolved alerts for these clusters, grouped by severity.

4. **Recent failed tasks**

Call `prism_execute` with the list-tasks operation, `_filter="status eq 'FAILED'"` and a recent time window, to surface failed operations that alerting may not cover.

5. **Compile**

Per cluster: node summary, critical/warning alert counts with the top items, failed task list, and an overall verdict (healthy / attention / critical).

Parameters

| Parameter | Type | Required | Description | |-----------|------|----------|-------------| | cluster_name | string | No | Check a single cluster instead of all |

Examples

/cluster-health
/cluster-health --cluster_name "prod-east"

Error Handling

**Multiple calls per cluster:** On large estates, summarize alert counts rather than fetching every alert body
**Missing namespace tools:** Report which of `clustermgmt` / `monitoring` / `prism` is unavailable instead of presenting a partial check as complete