How ServiceMon Prevents Downtime — Proactive Alerts & Automated Recovery

Getting Started with ServiceMon: Installation, Dashboards, and Best Practices

1. Quick overview

ServiceMon is a server and service monitoring tool that collects metrics, sends alerts, and offers dashboards for observability. The rest of this guide assumes a default single-server setup and an on-prem or cloud VM.

2. Installation (single-node Linux)

1. Prerequisites

   • Linux (Ubuntu 20.04+ or CentOS 7+)
   • Docker 20.10+ and Docker Compose 1.29+ (recommended)
   • 2 GB RAM, 2 CPU, 10 GB disk free

2. Install via Docker Compose

   • Create a directory and a docker-compose.yml with required services: service-mon, database (Postgres), and optionally alerting (Prometheus Alertmanager) and reverse proxy (nginx).
   • Example compose snippet:

     Code
     Copy CodeCopied
     version: ‘3.7’ services:servicemon:
     
      
     image: servicemon/latest ports: ['8080:8080'] environment:   - DB_HOST=db   - DB_USER=sm   - DB_PASS=changeme depends_on: ['db'] 
     db:
     image: postgres:13 environment:   - POSTGRES_USER=sm   - POSTGRES_PASSWORD=changeme   - POSTGRES_DB=servicemon 
     
     
     

   • Start: docker compose up -d

3. Alternative: Package manager

   • Use the provided .deb/.rpm if preferred: sudo dpkg -i servicemon_X.Y.Z.deb or sudo rpm -i servicemon-X.Y.Z.rpm.

4. Initial setup

   • Visit http://:8080 and complete the web installer: create admin user, configure DB (if not using default), set base URL, and enable SMTP for alerts.

3. Dashboards

• Default dashboards
  • Overview: system health, active incidents, alert status.
  • Host detail: CPU, memory, disk, network, and process metrics.
  • Service dependency: shows upstream/downstream relationships and impact.
• Custom dashboards
  • Use the built-in query editor to add panels for specific metrics (e.g., request latency p95, error rate).
  • Recommended panels: CPU load (1m/5m), memory used (%), disk I/O, network throughput, request latency histograms, and error rate trend.
• Best practices for dashboards
  • Keep overview dashboards high-level (no more than 8-10 panels).
  • Use one dashboard per team or service domain.
  • Use color thresholds for immediate signal (green/yellow/red).
  • Annotate incidents and deploys to correlate spikes with events.

4. Alerting & Notification

• Alert rules
  • Define rules for key signals: high CPU, memory pressure, disk near full, service unresponsive, error rate spike, latency SLO breaches.
  • Use short evaluation windows for fast failures (e.g., 1–5m) and longer windows for noisy metrics (e.g., 15m).
• Notification channels
  • Configure email, Slack, PagerDuty, and webhook integrations.
  • Use escalation policies and grouped alerts to avoid paging for transient issues.
• Reduce noise
  • Use alert grouping, mute windows (maintenance), and deduplication.
  • Add suppressions for known deploy-related alerts.

5. Best Practices

• Instrument services properly
  • Expose metrics in Prometheus-compatible format or use the provided agent.
  • Tag metrics with service, environment, and instance identifiers.
• SLO-driven alerting
  • Define SLOs and create alerts tied to error budget burn rather than raw thresholds.
• Secure the deployment
  • Enable HTTPS (letsencrypt/nginx), use strong admin passwords, rotate API keys, and restrict access via VPN or IP allowlist.
  • Encrypt database credentials and enable DB backups.
• Scalability
  • For production, separate components: dedicated DB, horizontal collectors, and replicated frontends behind a load balancer.
  • Use sharding/partitioning for long-term metrics retention and configure retention policies.
• Operational hygiene
  • Regularly test alerting channels and run playbook runbooks for common incidents.
  • Keep the system and agents