What is the TICK stack, how do its parts map to Prometheus, and should you run Telegraf and the Node Exporter side by side?
TICK is Telegraf, InfluxDB, Chronograf and Kapacitor, a competing metrics stack. The two are compatible, but running Telegraf and the Node Exporter for the same basic metrics just collects everything twice. Pick one for the basics.
| TICK | Prometheus equivalent |
|---|---|
| Telegraf (collector agent) | Node Exporter |
| InfluxDB (TSDB) | Prometheus server |
| Chronograf + Kapacitor (UI, alerting) | Alertmanager functions |
Because both projects are open source, their developers made the tools interoperable. Telegraf can expose its data in Prometheus format via its Prometheus output plug-in, so Prometheus can scrape Telegraf instead of the Node Exporter.
Why many admins prefer Telegraf: it has many more plug-ins (Postfix, OpenLDAP, nginx, MySQL, Redis), and if a plug-in is missing, a shell script with the right output is enough. Extending the Node Exporter needs a custom plug-in, which is real work. If both run, configure each to collect only what the other cannot.
Go deeper:
influxdata/telegraf on GitHub — the agent and its plug-in catalogue.
Telegraf — Prometheus client output plug-in — how Telegraf exposes its data for Prometheus to scrape.- InfluxDB — Wikipedia — the TSDB at the heart of TICK.