Initial varys-monitor agent

HTTP service reachability checks for 12 services + agent watchdog.
Writes status dashboard to /opt/sites/varys/ and last-run.json for Citadel.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Dockerfile

@@ -0,0 +1,6 @@
+FROM python:3.12-slim
+WORKDIR /app
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+COPY main.py config.yaml ./
+CMD ["python", "main.py"]

config.yaml

@@ -0,0 +1,36 @@
+# Services varys checks for HTTP reachability.
+# Any HTTP response (including 4xx) = UP. Connection error/timeout = DOWN.
+# Add/remove entries freely — no code changes needed.
+
+services:
+  # Agent infrastructure (on proxy Docker network — reachable by service name)
+  - name: hodor-gateway
+    url: http://hodor-gateway:8200/health
+  - name: citadel-mcp
+    url: http://citadel-mcp:8300/sse
+
+  # Core services
+  - name: open-webui
+    url: http://172.27.40.3:3010
+  - name: nginx-proxy-manager
+    url: http://172.27.40.3:81
+  - name: portainer
+    url: https://172.27.40.3:9443
+  - name: gitea
+    url: http://172.27.40.3:3000
+  - name: uptime-kuma
+    url: http://172.27.40.3:3002
+  - name: headscale
+    url: http://172.27.40.3:8080
+  - name: vaultwarden
+    url: http://172.27.40.3:8222
+  - name: plane
+    url: http://172.27.40.3:8095
+  - name: homarr
+    url: http://172.27.40.3:7575
+  - name: netbox
+    url: http://172.27.40.3:8100
+
+# One-shot agents to watchdog (reads /opt/agent-os/logs/<name>/last-run.json).
+agents:
+  - name: bran-changelog

main.py

@@ -0,0 +1,164 @@
+import json
+import logging
+import os
+import time
+from datetime import datetime, timezone
+from pathlib import Path
+
+import httpx
+import yaml
+
+logging.basicConfig(level=logging.INFO, format="%(asctime)s [varys] %(message)s")
+logger = logging.getLogger("varys")
+
+SITES_DIR = Path(os.getenv("SITES_DIR", "/opt/sites"))
+AGENT_OS_DIR = Path(os.getenv("AGENT_OS_DIR", "/opt/agent-os"))
+CONFIG_FILE = Path(__file__).parent / "config.yaml"
+
+
+def check_service(name: str, url: str) -> dict:
+    start = time.monotonic()
+    try:
+        # stream=True so we get the status code without reading the body (handles SSE endpoints)
+        with httpx.stream("GET", url, timeout=5, verify=False, follow_redirects=True) as r:
+            ms = int((time.monotonic() - start) * 1000)
+            status = "up" if r.status_code < 500 else "degraded"
+            return {"name": name, "status": status, "code": r.status_code, "ms": ms}
+    except Exception:
+        ms = int((time.monotonic() - start) * 1000)
+        return {"name": name, "status": "down", "code": None, "ms": ms}
+
+
+def read_agent_status(name: str) -> dict:
+    path = AGENT_OS_DIR / "logs" / name / "last-run.json"
+    if not path.exists():
+        return {"name": name, "status": "unknown", "timestamp": None, "result": "no data"}
+    try:
+        data = json.loads(path.read_text())
+        return {
+            "name": name,
+            "status": data.get("status", "unknown"),
+            "timestamp": data.get("timestamp", ""),
+            "result": data.get("result", ""),
+        }
+    except Exception as e:
+        return {"name": name, "status": "error", "timestamp": None, "result": str(e)}
+
+
+def render_html(services: list, agents: list) -> str:
+    now = datetime.now().strftime("%Y-%m-%d %H:%M")
+
+    SERVICE_COLOURS = {"up": "#3fb950", "degraded": "#d29922", "down": "#f85149"}
+    AGENT_COLOURS = {"success": "#3fb950", "failure": "#f85149"}
+
+    service_cards = ""
+    for s in services:
+        colour = SERVICE_COLOURS.get(s["status"], "#8b949e")
+        label = s["status"].upper()
+        ms_text = f"{s['ms']} ms" if s["status"] != "down" else "—"
+        service_cards += f"""
+    <div class="svc">
+      <span class="dot" style="background:{colour}"></span>
+      <span class="svc-name">{s['name']}</span>
+      <span class="svc-label" style="color:{colour}">{label}</span>
+      <span class="svc-ms">{ms_text}</span>
+    </div>"""
+
+    agent_rows = ""
+    for a in agents:
+        colour = AGENT_COLOURS.get(a["status"], "#8b949e")
+        ts = a["timestamp"][:16].replace("T", " ") + " UTC" if a["timestamp"] else "—"
+        agent_rows += f"""
+      <tr>
+        <td><span class="dot" style="background:{colour}"></span>{a['name']}</td>
+        <td style="color:{colour}">{a['status'].upper()}</td>
+        <td>{ts}</td>
+        <td class="dim">{a['result']}</td>
+      </tr>"""
+
+    return f"""<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>Varys — Status</title>
+  <style>
+    * {{ box-sizing: border-box; margin: 0; padding: 0; }}
+    body {{ font-family: monospace; background: #0d1117; color: #c9d1d9; padding: 2rem; }}
+    h1 {{ color: #58a6ff; margin-bottom: 0.25rem; }}
+    h2 {{ color: #8b949e; font-size: 0.85rem; text-transform: uppercase; letter-spacing: 0.05em; margin: 2rem 0 0.75rem; }}
+    .meta {{ color: #8b949e; font-size: 0.85rem; margin-bottom: 2rem; }}
+    .meta a {{ color: #8b949e; }}
+    .dot {{ display: inline-block; width: 8px; height: 8px; border-radius: 50%; margin-right: 6px; vertical-align: middle; }}
+    .svc {{ display: flex; align-items: center; gap: 0.75rem; border: 1px solid #30363d; border-radius: 6px; padding: 0.6rem 1rem; margin-bottom: 0.5rem; }}
+    .svc-name {{ flex: 1; }}
+    .svc-label {{ font-size: 0.8rem; min-width: 5rem; }}
+    .svc-ms {{ color: #8b949e; font-size: 0.8rem; min-width: 4rem; text-align: right; }}
+    table {{ width: 100%; border-collapse: collapse; }}
+    td {{ padding: 0.6rem 1rem; border-bottom: 1px solid #21262d; font-size: 0.9rem; }}
+    .dim {{ color: #8b949e; }}
+  </style>
+</head>
+<body>
+  <h1>Varys — Status</h1>
+  <p class="meta">Updated {now} &nbsp;·&nbsp; <a href="/">← home</a></p>
+  <h2>Services</h2>
+  {service_cards}
+  <h2>Agents</h2>
+  <table><tbody>{agent_rows}</tbody></table>
+</body>
+</html>"""
+
+
+def render_summary_md(services: list, agents: list) -> str:
+    now = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M UTC")
+    up = sum(1 for s in services if s["status"] == "up")
+    down = [s["name"] for s in services if s["status"] in ("down", "degraded")]
+
+    lines = [f"# Varys — Status Summary\n\nUpdated: {now}\n"]
+    lines.append(f"## Services: {up}/{len(services)} up")
+    if down:
+        lines.append(f"Issues: {', '.join(down)}")
+    lines.append("")
+    lines.append("## Agents")
+    for a in agents:
+        ts = a["timestamp"][:16].replace("T", " ") + " UTC" if a["timestamp"] else "no data"
+        lines.append(f"- {a['name']}: {a['status']} ({ts}) — {a['result']}")
+    return "\n".join(lines)
+
+
+def main():
+    config = yaml.safe_load(CONFIG_FILE.read_text())
+
+    services = []
+    for svc in config.get("services", []):
+        result = check_service(svc["name"], svc["url"])
+        logger.info(f"{svc['name']}: {result['status']} ({result['ms']} ms)")
+        services.append(result)
+
+    agents = []
+    for agent in config.get("agents", []):
+        result = read_agent_status(agent["name"])
+        logger.info(f"agent {agent['name']}: {result['status']}")
+        agents.append(result)
+
+    out_dir = SITES_DIR / "varys"
+    out_dir.mkdir(parents=True, exist_ok=True)
+    (out_dir / "index.html").write_text(render_html(services, agents))
+    (out_dir / "last-output.md").write_text(render_summary_md(services, agents))
+
+    down_count = sum(1 for s in services if s["status"] != "up")
+    status = {
+        "agent": "varys-monitor",
+        "timestamp": datetime.now(timezone.utc).isoformat(),
+        "status": "success",
+        "result": f"{len(services)} services checked, {down_count} not up",
+    }
+    log_dir = AGENT_OS_DIR / "logs" / "varys-monitor"
+    log_dir.mkdir(parents=True, exist_ok=True)
+    (log_dir / "last-run.json").write_text(json.dumps(status, indent=2))
+    logger.info(f"done — {len(services)} services, {down_count} not up")
+
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -0,0 +1,2 @@
+httpx
+pyyaml

run.sh

@@ -0,0 +1,15 @@
+#!/usr/bin/env bash
+# Run varys-monitor: builds image, checks all services and agents, exits.
+# Schedule via cron: */15 * * * * /opt/stacks/varys-monitor/run.sh >> /opt/agent-os/logs/varys.log 2>&1
+
+set -e
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+docker build -q -t varys-monitor "$SCRIPT_DIR"
+docker run --rm \
+  --network proxy \
+  --volume /opt/sites:/opt/sites \
+  --volume /opt/agent-os:/opt/agent-os \
+  --env SITES_DIR=/opt/sites \
+  --env AGENT_OS_DIR=/opt/agent-os \
+  varys-monitor