client_python VS node_exporter

Here you go: https://stackoverflow.com/questions/75652326/celery-spawn-si...

Plus some adjacent discussion on GitHub: https://github.com/prometheus/client_python/issues/902

Hope that helps!

Prometheus, which is also a CNCF open source project, collects metrics data by scraping HTTP endpoints and then stores that data in a time series database that uses a multidimensional model. It’s a powerful tool for gathering metrics about your application and it also includes alerting functionality that you can use to notify your teams when issues come up. Prometheus includes a client library for Python.

We have a java one but the principle is the same. Install the Prometheus client ( https://github.com/prometheus/client_python) ,create the metrics you want, then push jmx settings to Prometheus.

The code above is copied from the official documentation of prometheus_client which simply creates a new metric named request_processing_seconds that measures the time spent on that particular request. We'll cover other types of metrics later in this post.

Just use the client? https://github.com/prometheus/client_python

I've experimented with the official Prometheus python client, i really really like the way they use decorators to instrument. I've tried to measure latency with multiple types of metrics (histogram, & summary), i see the value in both of them, but the one that between fits my objective is the histogram metric type. Great!

Now is a good time to expand your event loop by adding metrics collection of the event handler functions and also use that endpoint as a liveness probe. E.g. https://github.com/prometheus/client_python just add the event handled, success/error and the duration as a histogram (look for examples of tracking http requests served)

I've spent along time trying to figure out a bug that I'm facing while using Prometheus from its python client.

The official Prometheus Python client library makes this easy, no need to worry about the export file format.

$ wget https://github.com/prometheus/node_exporter/releases/download/v1.7.0/node_exporter-1.7.0.linux-amd64.tar.gz $ tar -xzvf node_exporter-1.7.0.linux-amd64.tar.gz

Node Exporter

The Prometheus node_exporter can provide this information, doesn't require root. You could run it as a systemd user unit if you don't have root.

I personally find it best to learn through experimentation. Start with reading a bit about Prometheus and Grafana through their docs, and then familiarise yourself with setting up a local Prometheus + Grafana instance either locally or with docker using docker-compose, along with something to generate /metrics endpoint for Prometheus to scrape from such as a custom Prometheus exporter in Python or using Node Exporter.

Your best bet might be to fork node_exporter to get you more verbose socket stats: https://github.com/prometheus/node_exporter/blob/master/collector/sockstat_linux.go

I use Grafana + Prometheus + Node Exporter.

Crazy idea, can't NX devices run Docker? I wonder if the node_exporter would work.

Kubernetes itself will not notify you, the way I've seen people do this, is to use something like kube-state-metrics or node_exporter, export that to Prometheus (or preferrably VictoriaMetrics because Prometheus is terrible IMO), and then setup alarms on that with alertmanager or equivalent, or just look at dashboards regularly with Grafana. Realistically I recommend only setting alerts on disk usage and application/database latency. CPU and memory utilization isn't a great metric to alert on a lot of the time.

You may setup node exporter and collect metrics with prometheus for example. Its not quite "simple" way, but still you may find it useful.