EMZETT.
Login

Emzett Wiki

IT monitoring and availability: Nagios, Zabbix and uptime monitoring

Properties1
tagsglossar, it-betrieb, monitoring

2 min read

IT monitoring and availability: Nagios, Zabbix and uptime monitoring

In short: Monitoring continuously checks whether services and servers are reachable and healthy and alerts on problems; availability is measured in percent and managed via thresholds and alarms.

In more detail: “We notice when customers call” is too late. Monitoring detects outages and trends (disk filling up, memory short, certificate expiring) before damage occurs.

In Depth

What is monitored

  • Reachability (ping/HTTP/TCP port), response time.
  • System values: CPU, RAM, disk space, I/O, network throughput, temperature.
  • Services/processes, database, queues, logs/error rates.
  • Certificates (expiry), domain, backups (do they run?), UPS, printers/toner (SNMP).
  • Application (APM): requests, errors, latencies, business metrics (orders per minute).
  • Synthetic monitoring: simulated user goes through login/purchase.

Tools

  • Nagios/Icinga/Checkmk: classic check-based monitoring.
  • Zabbix: agent, SNMP, auto-discovery, dashboards.
  • Prometheus + Grafana: metrics, Alertmanager (cloud-native).
  • Uptime Kuma (self-hosted, simple), UptimeRobot, Better Stack, Pingdom: external HTTP checks, status pages.
  • ELK/Loki (logs), Sentry (errors), OpenTelemetry (traces).
  • A status page (like /status) informs users honestly about disruptions.

Alarms and availability

  • Thresholds with warning (80 %) and critical (95 %); hysteresis and retries (alert only after 3 failures) against false alarms.
  • Avoid alert fatigue: only actionable alarms, priorities, on-call schedule, runbooks (what to do on alarm X?).
  • Channels: email, chat, push, SMS, phone for critical.
  • Availability = (total time − outage) / total time; 99.9 % ≈ 43 min outage/month; set SLO/SLA; error budget.
  • Four golden signals: latency, traffic, errors, saturation.

Getting started and exam

  • Start small: Uptime Kuma for websites + certificate + disk-space alarm; notification to phone.
  • Root cause analysis: dashboard + logs + history (what changed? deployment?).
  • Capacity planning from trends.
  • Privacy: logs may contain personal data (IP), limit retention.
  • Upkeep: keep checks current, test alarms, monitor the monitoring (dead man’s switch).
  • IHK: calculate availability, explain monitoring terms (SNMP, agent, polling vs. trap, threshold).

See also: Prometheus and metrics, Dashboards with Grafana, SNMP, IT contingency planning

Backlinks