Highlights on Prometheus Day 2022 EU (Valencia)
Last week in Valencia, during the KubeCon EU Cloud Native Collocated Events, I had the chance to attend to the Prometheus Day Europe.
It was a great opportunity to get back in contact with the amazing Prometheus open source community, contributors and maintainers, as well as sharing knowledge and ideas in and after talks. And of course it was great having some networking time.
In case you couldn’t attend, here are some of the things I enjoyed the most!
Prometheus open source community
The day before the main event, there was a shared Project Meeting of the Prometheus and Thanos projects. In this meeting there were different technical and organizational discussions. Some that I found particularly interesting, were the proposal to improve and unify the feature requests in Prometheus projects and how they are accepted, and the creation of a Governance Working Group.
You can read the full notes of this meeting and previous ones in the public Prometheus dev summit document.
Love for Ukraine
The Prometheus team reached the consensus to “strongly condemn Russia’s illegal invasion of Ukraine”. Please, don’t miss the moving tribute that Richard Hartmann, the Community Director, paid to the colleagues and members of the Prometheus community that couldn’t travel because of the conflict. I honestly hope they can join us for the next edition.
Sessions and lighting talks
There were some hot topics during the day that appeared in several slide decks.
Björn Rabenstein made an inspiring talk about the convenience of using anomaly detection for alerting. I also gave my two cents in How to start implementing anomaly detection in PromQL, a lighting talk I had the pleasure to give.
As more companies adopt monitoring solutions, their needs, and expectations also grow and reach the point of asking about AIOps. These two talks are a starting point for both setting the correct expectations and to start exploring the first models and results.
Exemplars in Prometheus
Exemplars in Prometheus was also another hot topic. Ramon Guiu and Ian Billett made both presentations, talking about how Prometheus implements the support for exemplars and how to use them in real world applications. Adding exemplars to the Prometheus metrics allows us to combine two of the main pillars of observability: metrics and tracing, giving the chance to create connections between the monitoring platforms and tracing applications.
Prometheus Internals and the Operator
I also had the chance to assist with these talks that explained the internals of some components of Prometheus. Jayapriya Pai and Haoyu Sun made a deep dive on the architecture and functioning of the Prometheus Operator. Aditi Ahuja shared her learning path of Prometheus instrumentation and explained some great tips about how to implement tests.
Also, Harkishen Singh explained how Prometheus stores data in blocks and manages the indexing for queries.
The serverless and FaaS world also had its time during the Prometheus Day. Bartłomiej Płotka and Saswata Mukherjee gave a talk where they shared knowledge and experiences on the challenges to monitor ephemeral processes. Likewise, Alex Ellis explained his insights on how they implemented autoscaling using Prometheus metrics in OpenFaaS.
Some gaming too
Finally, my colleague David Lorite explained to everyone how he used Prometheus to monitor the video game Counter Strike Global Operations, not only the players per level, but also how to rightsize your server to optimize costs.
It was so great traveling to Valencia to attend all the KubeCon events, and especially Prometheus Day. It was my first on-site event in a very long time, and I enjoyed it a lot. In this article, I showed you the talks I could attend or found most interesting.
There were other amazing talks that you can see in this YouTube playlist and find the slides in the event Sched page. If you have thoughts, or want to point something out about the article or the Prometheus Day, you can ping me on Twitter!