Need help in exporting the top dashboard

Help
,
1

Problem / Question

I’m trying to understand whether the data shown in the Top dashboard (especially per-process CPU usage) is exportable, and if so, whether it can be exported via Prometheus.

Additionally, the main purpose I’m looking for is to visualize process-level CPU usage over time (not just a point-in-time snapshot). I’d like to confirm whether this capability already exists in Netdata (either via built-in dashboards or exported metrics), or if the Top dashboard is primarily meant for real-time inspection only.

Specifically:

  • Is the data backing the Top dashboard exposed as time-series metrics?

  • Can this data be exported via the Prometheus exporter?

  • If process CPU usage over time is already supported, which dashboard(s) or metrics should be used?

Relevant docs followed / actions taken

  • Looked through Netdata dashboards and explored the Top / Processes views

  • Reviewed exporter-related documentation to understand Prometheus support

What I expected to happen

I expected that the data powering the Top dashboard (process CPU usage) would either:

  • Be available as historical time-series metrics, and

  • Be exportable via Prometheus so it can be visualized externally (e.g., Grafana) over longer time windows.

If the Top dashboard is intentionally real-time only, I’d like to understand what the recommended way is to track process CPU usage trends over time in Netdata.

2

Hi, @Devansh_Wassista. All collected metrics can be scraped in Prometheus format from http://IP:PORT/api/v1/allmetrics?format=prometheus. Metrics related to top processes use the netdata_app_* prefix.

3

Hi @ilyam8 Got it.

Using netdata_app_* metrics I can chart process CPU/memory over time, but they are grouped by app_group. In my setup, many processes fall under generic groups like bash, which isn’t very useful.

Is there a way to export Top dashboard–like CPU usage broken down by PID, command/cmdline, or a more granular selector similar to the Top dashboard checkboxes?

For example, currently I see metrics like:

netdata_app_mem_usage_MiB_average{app_group="ModemManager", ...}

but I’m looking to track specific processes over time, not just grouped apps.

Thanks a lot.

4

Is there a way to export Top dashboard–like CPU usage broken down by PID

No, there is no way to do that. Netdata doesn’t store this info.

5

I’m still a bit confused though — since the Top dashboard is able to display per-PID / per-process(cmdline) CPU usage in real time, I assume it would be exportable somehow.

image
image1591×362 35.1 KB

6

This information is kept in memory only. Since Netdata doesn’t store it in the database, it isn’t available for export.

7

Got it, thanks for the clarification!