store: Fix: Server Health Monitor n8n workflow - SSH auth and script paths

graph/fixes/fix-server-health-monitor-n8n-workflow-ssh-auth-and-script-p-d5822f.md

@@ -0,0 +1,35 @@
+---
+id: d5822f92-ef0f-4f31-8fb0-37aaf6dfbf29
+type: fix
+title: "Fix: Server Health Monitor n8n workflow - SSH auth and script paths"
+tags: [n8n, homelab, monitoring, claude-runner, ssh, fix]
+importance: 0.8
+confidence: 0.8
+created: "2026-02-20T04:16:54.750680+00:00"
+updated: "2026-02-20T04:16:54.750680+00:00"
+---
+
+# Server Health Monitor n8n Workflow Fix
+
+## Workflow
+- Name: "Server Health Monitor - Claude Code"
+- ID: `p7XmW23SgCs3hEkY`
+- n8n instance: http://10.10.0.210:5678
+
+## Root Causes
+
+1. **SSH credential pointed at wrong host** — needed updating to `10.10.0.148` (CT 302 claude-runner) with `n8n_runner_key` (ed25519 key type). Credential must be updated manually in n8n UI (Credentials → SSH Private Key account) since the API cannot modify credentials.
+
+2. **Wrong script path** — command was `/root/health-check.sh` which does not exist. Updated to:
+   ```
+   /root/.claude/skills/server-diagnostics/.venv/bin/python /root/.claude/skills/server-diagnostics/health_check.py
+   ```
+
+3. **Parse node rewritten** — updated to handle JSON output with `status`, `issues`, and `escalations` fields from health_check.py.
+
+4. **Discord Alert node updated** — refreshed to match new output format.
+
+5. **remediate.sh created** — wrapper script on CT 302 to handle escalation path (see separate memory).
+
+## Key Detail
+health_check.py exits with code 2 when escalation is needed, triggering the remediate.sh path in the workflow.