KubeCon + CloudNativeCon preview: Grafana Labs team members talk about Prometheus, Cortex, and OpenMetrics

Published: 14 Aug 2020

KubeCon + CloudNativeCon EU is going virtual next week, and Grafana Labs team members will be among the thousands of cloud native enthusiasts gathering online. It’s not too late to join us; you can still register here.

Here’s where you will find Grafana Labs team members during the conference:

Talks

Scaling Prometheus: How We Got Some Thanos into Cortex

Tuesday, August 18 @ 17:45 - 18:20 CEST Grafana Labs software engineer and Cortex maintainer Marco Pracucci joins forces with Hashicorp’s Thor Hansen to present the new experimental Cortex blocks storage, based on Thanos and Prometheus TSDB. The goal: to reduce the Cortex operational cost without compromising scalability and performance. The two will also share lessons learned running Cortex at scale.

Prometheus Deep Dive

Wednesday, August 19 @ 13:45 - 14:20 CEST Grafana Labs Software Engineer and Prometheus maintainer Goutham Veeramachaneni will team with Red Hat’s Bartlomiej Plotka to present advanced use cases of Prometheus. Attendees will also have the opportunity to have their questions fielded by Prometheus maintainers during the Q&A at the end.

Make Prometheus Use Less Memory and Restart Faster

Thursday, August 20 @ 13:00 - 13:35 CEST Grafana Labs Software Engineer Ganesh Vernekar, a maintainer of the Prometheus storage engine, will talk about the most common problem that causes Prometheus servers to run out of memory: too many time series in the head block, the part of the internal TSDB with the freshest data, which has to be kept in memory prior to consolidation into a block on disk. He’ll discuss the work that’s being done to persist parts of the head block earlier, thereby reducing both the memory footprint and the restart time.

What You Need to Know About OpenMetrics

Thursday, August 20 @ 13:45 - 14:20 CEST Grafana Labs Director of Community Richard “RichiH” Hartmann and Robust Perception’s Brian Brazil will lead an introduction to the OpenMetrics format for standardizing metric exposition, which simplifies how both developers and operators monitor their systems. They’ll answer FAQ including whether you need to rewrite your existing instrumentation, what’s needed to transition, and whether it will be supported by your monitoring system.

Better Histograms for Prometheus

Thursday, August 20 @ 14:30 - 15:05 CEST Grafana Labs' Björn “Beorn” Rabenstein first added robust histogram functionality to Prometheus in version 0.11.0 back in February 2015. Though it is simple and powerful and enables many important use cases, Rabenstein also finds it infuriatingly limited, mostly because of its high per-bucket cost. In this talk, he will share his latest findings from recent efforts to bring efficient high-resolution histograms to Prometheus.

In the Expo Hall

The Prometheus Booth will be located in the Project Pavilion in the Expo Hall. Goutham Veeramachaneni, Richard Hartmann, and Björn Rabenstein from Grafana Labs will be available to talk about what’s new in the Prometheus project.

Press and industry analysts looking to schedule meetings during the show can send an email to grafana@offleashpr.com.

We hope to see you (virtually) at KubeCon + CloudNativeCon EU!

Related Posts

In his KubeCon talk, Goutham Veeramachaneni offered advice about performance tuning and day 2 operations for Cortex clusters at scale.
At KubeCon this week, Tom Wilkie is giving two talks on Cortex. ICYMI, here’s the latest news about the technology used in Grafana Cloud’s Hosted Prometheus.
While prepping for large new tenants, the Cortex community has run into some challenges. But along the way we've found solutions that include compactor, a query-frontend, and lots of sharding.