Learn · Platform landscape · 📈
Zabbix, Prometheus, Grafana, ELK, Nagios, Datadog:what each is for.
Monitoring tells you something is down; observability tells you why. The open-source stack covers 90% of what Indian businesses need — the paid tools buy convenience and support.
| Tool | Type | Licence | Best at | Pair with |
|---|---|---|---|---|
| Zabbix | Infrastructure monitoring | Open source | Servers, network devices (SNMP), templates, alerting, SLA reports | Grafana for dashboards; Opsgenie/PagerDuty |
| Prometheus + Alertmanager | Metrics (pull, time-series) | Open source | Kubernetes, services, custom exporters | Grafana, Loki, Thanos for long-term |
| Grafana | Dashboards & alerting | Open source / cloud | Visualising anything (Prometheus, Zabbix, SQL, logs) | Everything above |
| Loki / ELK (Elasticsearch, Logstash, Kibana) / OpenSearch | Logs | Open source | Centralised logs, search, retention | Promtail/Filebeat agents |
| Nagios / Icinga / CheckMK | Classic checks | Open source / paid tiers | Simple host/service checks; CheckMK is the modern one | — |
| LibreNMS | Network monitoring | Open source | SNMP auto-discovery, ports, traffic graphs | Zabbix or Grafana |
| Uptime Kuma | Uptime checks | Open source | HTTP/TCP/ping checks with status page — tiny footprint | WhatsApp/Telegram alerts |
| Datadog / New Relic / Dynatrace | SaaS APM/observability | Paid (per host/GB) | Application tracing, all-in-one, fast setup | Budget: ₹5–25k/host/month |
| Opsgenie / PagerDuty | On-call & escalation | Paid | Who gets woken, escalation, schedules | Any alert source |
How to read this
- Our standard open stack: Zabbix (infra) + Prometheus/Grafana (services/K8s) + Loki (logs) + Uptime Kuma (external) + Opsgenie (on-call).
- Alert design matters more than tool choice: page only on symptoms customers feel; everything else is a ticket.
- Learn Zabbix and Prometheus/Grafana — both appear in most Indian NOC/DevOps job descriptions.