在 NiFi 中创建自定义 Prometheus 指标

Question

We are going to use the Prometheus framework to monitor our NiFi instance and dataflows inside.

In order to achieve this, we already configured PrometheusReportingTask to expose pre-existing NiFi metrics.

The amount of metrics is pretty full but we would like to create custom Prometheus metrics in or custom/predefined NiFi processors and expose them using PrometheusReportingTask.

Is it possible to implement it?

Thanks!

Answer 1

PrometheusReportingTask has hardcoded metrics registries and can't be extended as such. You could create your own ReportingTask that opens another port as a Prometheus scrape target and expose your metrics, then your Prometheus instance can scrape both targets.

If you don't want to code a full ReportingTask instance (and a NAR to keep it in), take a look at ScriptedReportingTask , you can put your custom code in there so you don't need to create a project/module/NAR for your custom reporting task.