Prometheus – Best practices and pitfalls with Julius Volz

Last year, we had the pleasure of hosting a webinar on Byline about “Prometheus – Best practices and pitfalls” by Julius Volz, co-founder of Prometheus and founder of PromLabs, which provides trainings for the most effective monitoring concepts with Prometheus.

The informative webinar covered various aspects of Prometheus that remain highly relevant today – including an in-depth discussion of the following topics:

  • What to consider when adding metrics to services (USE and RED method; metric and label naming; tracking of errors, successes, and totals; how to handle missing series; normalization of metrics)
  • What to pay attention to when creating alerting rules (general alerting guidelines; unhealthy or missing targets; “for”-durations; common and useful labels)
  • How to query properly with PromQL (scope selectors to jobs; functions for counters and gauges; rate () time windows)
  • Topology in monitoring – common problems in architectures (per-machine uber- exporter versus per-process exporter; abusing Pushgateway or federation in Prometheus)

As Prometheus experts ourselves, we can only highly recommend to re-watch this informative session below.

Thank you again to Julius for hosting this webinar session. Check out promlens.com if you want to learn more about PromLens, which was demoed at the end of the presentation, or visit training.promlabs.com for more trainings with Julius.

Finally, feel free to contact us at support@nexclipper.io if you have any questions related to the webinar or if you would like to connect with our team.