Nightingale is similar to Grafana in that it can integrate with multiple data sources. The most commonly used data source is Prometheus (other data sources that are Prometheus-API-compatible, such as VictoriaMetrics, Thanos, and M3DB, can all be considered Prometheus types), so the two are closely related.

If you have any of the following needs, Nightingale is worth considering:

• You have multiple time-series databases (Prometheus, VictoriaMetrics, etc.) and want a unified platform to manage all kinds of alert rules with permission controls.
• Prometheus’s alert engine is a single point — you worry about the single machine going down and the alert engine becoming unavailable.
• Besides Prometheus alerts, you also need alerts for ElasticSearch, Loki, ClickHouse, and other data sources.
• You need more flexible alert rule configuration, such as effective-time control, event relabeling, event-to-CMDB integration, and alert-triggered self-healing scripts.

Nightingale also has visualization capabilities similar to Grafana, but not as deep as Grafana. From my observations many companies adopt a combined approach (in the adult world, nothing is purely black or white; you take both):

• Data collection: a combination of various agents and exporters. Mainly Categraf (especially for machine monitoring, which integrates smoothly with Nightingale), supplemented with various Exporters.
• Storage: time-series storage mainly uses VictoriaMetrics, since VictoriaMetrics is Prometheus-compatible and offers better performance with a cluster edition. For most companies the single-node version is enough.
• Alert engine: Nightingale, convenient for cross-team management and collaboration. Comes with built-in out-of-the-box rules, very flexible alert configuration, and event Pipeline mechanism that makes integrating with your own CMDB easy.
• Visualization: Grafana — flashier charts, a much larger community, and tons of pre-built dashboards available from grafana.com, saving effort.
• Alert event on-call dispatch: Flashduty — integrates with Zabbix, Prometheus, Nightingale, various cloud monitoring services, Elastalert, etc., centralizing alert events on a single platform for unified noise reduction, scheduling, acknowledge/escalate, response, and dispatch.

FAQ

Q1: If I use Nightingale, do I still need to deploy Prometheus?

A: It depends on the scenario:

• Small scale + demo: use the VictoriaMetrics TSDB bundled with Nightingale — no Prometheus needed;
• Existing Prometheus: just use it directly as a data source for Nightingale;
• Large scale: we recommend horizontally scalable TSDBs like VictoriaMetrics / Thanos / Mimir instead of single-node Prometheus.

Q2: Can I migrate Prometheus alerting rules directly to Nightingale?

A: Yes. Nightingale supports parsing and importing Prometheus alerting rules files. Or use Alert Rule Templates for bulk import.

References

• Prometheus Data Source
• Alert Rules
• VictoriaMetrics Docs