It’s anniversary time, concluding the third year of OpenObservability Talks. And what a year it was. Last year we ran the show partially under the pandemic conditions, and this year we got back to meeting in person and to conferences. We even had a special episode live from the KubeCon event with the Cloud Native Computing Foundation (CNCF) leadership, in an in person fireside chat rather than the usual virtual format.
We also got a new look for the show beginning of 2023 (so half year yellow, the other half blue 🙂 )
This year we discussed observability in practice, covering challenges with monitoring Kubernetes and cloud native environments, and learning from end user practices at hyperscalers such as Google and Meta. As always, we covered the state of prominent projects such as Prometheus and OpenTelemetry in depth, alongside new projects such as Backstage and OpenCost.
Observability is needed in more places than ever. We discussed hot topics such as Platform Engineering and FinOps, and how observability plays a key role in these domains.
Let’s see what we’ve been chatting about.
The Cloud Native Computing Foundation (CNCF) Ecosystem
KubeCon Europe 2023 was a special one. I got nominated as a CNCF Ambassador, and the podcast got invited as a CNCF media partner. We streamed a special live episode from the show floor with the CNCF’s Head of Ecosystem, Taylor Dolezal, to report hot updates and insider insights, finally doing an in-person format.
I also had the privilege of hosting Chris Aniszczyk, the CTO of the CNCF, to discuss the landscape and evolution of the CNCF, trends in observability, and open source in general, including some fascinating predictions.
OpenTelemetry in Practice
Those who follow me know that OpenTelemetry is a passion of mine. Apparently, it’s not just me. The top episode of this year was the one dedicated to the state of OpenTelemetry, with hot updates from the KubeCon 2022 conference. I was joined by Alolita Sharma, co-chair of the CNCF Technical Advisory Group for Observability and a member of the OpenTelemetry Governance Committee, among her many titles.
I also devoted an episode to the intricate topic of application instrumentation, this magic of getting your application to emit telemetry data. I invited Eden Federman, Co-Founder & CTO at keyval and the creator of the Go automatic instrumentation (now part of OpenTelemetry), to learn what options are available in different programming languages, such as Java, Python and Go, and what OpenTelemetry offers in this domain.
Monitoring Kubernetes and Prometheus updates
Prometheus is another prominent CNCF observability project I cover regularly on the show, and with major features released, I invited Julien Pivotto, a maintainer of Prometheus, to hear it right from the source.
Despite advancements in the monitoring realm, many experience ongoing challenges in monitoring Kubernetes environments. High churn rate of pod metrics, proliferation of metrics with low usage, and configuration complexity are some of the issues I discussed with Aliaksandr Valialkin, CTO at VictoriaMetrics.
Platform Engineering and Developer Platforms
Platform Engineering is gaining a lot of attention, and I turned to someone who’s done it for some time at scale. With George Hantzaras, Director of Cloud Platform Engineering at Citrix, I discussed the concept of Platform Engineering, its relationship with DevOps and SRE, and the crucial role observability plays in building and maintaining robust platform engineering practices.
I also wanted to learn how Spotify’s Platform team does it internally, and how they came up with the Backstage project, now a popular Internal Developer Platform open source project under the CNCF.
FinOps and Observability
As the economy puts strains, I delved into the challenges organizations face in monitoring the costs of their Kubernetes workloads and the FinOps practices and FinOps Foundation building around this discipline. I also explored OpenCost, a vendor-neutral open source project for cost monitoring and allocation, with Matt Ray, Senior Community Manager for the OpenCost project.
I also invited Ben Sigelman, co-founder and GM of Lightstep, to address the challenges associated with the cost of observability and the cardinality problem with today’s metrics. Ben was part of the observability team inside Google, and he shared some of the learnings from the hyperscaler front.
Observability at Hyperscalers
In addition to the Google insider experience from Sigelman, I also invited David Ostrovsky to learn how observability is treated at Meta, where a data-driven approach permeates various aspects of the company. We explored how observability is approached as a data analytics problem, providing insights that can be implemented by organizations of any scale.
New Year, New Ideas
The opening episode of the 4th year will be with another hyperscaler — eBay. I will host Vijay Samuel, Observability Architect at eBay, to learn how they run their planet-scale observability, and how they use an open source stack at that scale.
We will also discuss a new CNCF working group for defining a standard query language for observability, which Vijay is one of its initiators.
The episode will live-stream June 8 on YouTube and Twitch, and will publish to the podcast apps last Thursday of the month, as always. After the live date, you can also watch the replay here:
We’re working on more exciting things for the upcoming season, so stay tuned! The show is available on all the major podcast apps. And join the live streams on YouTube and Twitch to chime in with their questions and comments on chat. Follow us on Twitter @openoberv to stay up to speed on the live stream times, and to share your thoughts and comments on the show.
And thanks for your continued participation and support.
-Dotan Horovits @horovits