Last updated: September 1, 2025
Code Red Newsletter #5
Hi there,
Welcome back to your bi-weekly dose of observability geekery! With KubeCon + CloudNativeCon Europe 2025 just around the corner in London, it’s the perfect time to call out some must-see sessions for anyone who thrives on data, telemetry, and the sweet smell of well-instrumented code.
In Focus: KubeCon + CloudNativeCon Europe
As one of your friendly neighborhood KubeCon + CloudNativeCon Co-Chairs, I’ve spent months helping to shape an epic lineup of talks that push how we think about observability, AI, and platform engineering.
This newsletter is your cheat sheet to five sessions you don’t want to miss. But before we dive into the recommendations - a shameless plug for my own keynote, happening during Wednesday's keynotes: “The Observability Platform Engineering Advantage: From Zero-Code to Monitoring as Code”, where I’ll demo the OpenTelemetry Operator’s auto-instrumentation and Perses for monitoring as code.
While you’re at it, check out Michele Mancioppi’s talk, “The Art and Craft of No-touch Instrumentation”, on Tuesday, April 1 at Observability Day (a co-located KubeCon event).
If you’re around, make sure to swing by the Dash0 booth, S360. Book a live onsite demo at KubeCon and get a personalized walkthrough of how Dash0 simplifies observability and a chance to walk away with a limited-edition Dash0 backpack.
Keynote: Into the Black Box: Observability in the Age of LLMs
Building on LLMs is easy - until it’s not. They’re unpredictable, they hallucinate, and suddenly, your app is confidently wrong in ways you can’t even explain. Christine Yen will break down how to keep AI-powered systems observable, accountable, and sane - all while making sure your AI-powered product doesn’t turn into an expensive guessing machine.
Speaker: Christine Yen, CEO and Co-founder, Honeycomb
When: Wednesday, April 2 - 09:11 - 09:26 at the ICC Auditorium (level 0)
Link: https://sched.co/1txBR
Keynote: AI Enabled Observability Explainers - We Actually Did Something With AI!
I get it - another AI talk? Not this time! This one is for the skeptics. Instead of just plugging an LLM into everything and hoping for the best, eBay has actually cracked the code on making AI-driven observability predictable and accurate. Their "Explainers" don’t just analyze traces - they understand them. If you want to see how AI can genuinely help debugging, troubleshooting, and making sense of the madness, this is the talk to attend.
Speaker: Vijay Samuel, Principal MTS, Architect, eBay
When: Wednesday, April 2 - 09:33 - 09:48 at the ICC Auditorium (level 0)
Link: https://sched.co/1txBX
OTel Sucks (But Also Rocks!)
OpenTelemetry is both amazing and frustrating - and this talk will hilariously walk you through why. Expect unfiltered takes on SDK configuration nightmares, collector growing pains, and semantic conventions that feel like a necessary evil. But, despite all of this, OpenTelemetry remains the future. Join for candid insights, real-world war stories, and maybe a few rants - because let’s be honest, OTel is definitely here to stay.
Speakers: Juraci Paixão Kröhling, OllyGarden & Daniel Dyla, Dynatrace
When: Thursday, April 3 - 11:00 - 11:30 at Hall Entrance N10 - Room G (level 1)
Link: https://sched.co/1txHm
How To Adopt OpenTelemetry in an Enterprise Where Incumbent Vendor Tools Reign Supreme
Stuck in vendor lock-in hell? Wondering how to break free without breaking everything? Chris Weldon shares a three-year journey of migrating an entire enterprise to OpenTelemetry—while keeping execs, engineers, and the finance team happy.
Expect hard truths, real strategies, and inspiration for making OpenTelemetry work in even the most legacy-laden environments.
Speaker: Chris Weldon, Wolters Kluwer
When: Thursday, April 3 - 14:15 - 14:45 at Hall Entrance N10 - Room G (level 1)
Link: https://sched.co/1txHs
Observability Pipeline Query Languages: Present and the Future
Splunk, LogQL, ES/QL, OpenSearch PPL - why do we have so many query languages? And which one should you actually use? This talk takes you on a tour of observability pipeline query languages, their strengths, their quirks, and the potential for a more unified future. If you care about logs, traces, metrics, and making sense of them efficiently, this one’s for you.
Speaker: Jacek Migdal, Quesma
When: Thursday, April 3 - 16:45 - 17:15 at Hall Entrance N10 - Room G (level 1)
Link: https://sched.co/1txI1
I wouldn’t dare cheat you out of some interesting observability news...
Your flight or train ride shouldn’t be boring, so here are a few more nuggets from the world of observability - and a podcast to keep you company.
Introducing SIFT: The Next Stage of Observability
Troubleshooting shouldn’t feel like searching a needle in a haystack. That's why, at Dash0, we have introduced SIFT. SIFT stands for: Spam removal, Improve telemetry, Filtering & grouping, and Triage. It’s all about keeping your data consistent, structured, and easy to work with.
In Dash0, spam filters remove junk telemetry before it clogs up your system. Log AI, resource-centricity and semantic convention upgrades (plus more to come soon!) improve the telemetry you keep by standardizing and enriching it, makes filtering and grouping effortless so you can quickly spot patterns, and uses Triage to pinpoint issues without endless searching. Observability should be simple, and Dash0 takes on the complexity of making it so.
Code RED Podcast: Observability 2.0
Christine Yen, co-founder and CEO of Honeycomb, sat down with Mirko Novakovic, CEO of Dash0, to dive into the evolution of observability and why traditional monitoring wasn’t cutting it.
Instead of collecting more data, Honeycomb built its own database to make telemetry queryable, explorable, and actually useful for developers and ops alike.
It all ties back to the core of observability - understanding what’s happening in your system without guesswork.
Listen to the full episode here.
AI Agent Observability - Evolving Standards and Best Practices
AI agents aren’t just chatbots on steroids - they’re autonomous systems making real-time decisions. But how do you observe something that constantly evolves? Traditional monitoring falls short, so OpenTelemetry is setting new semantic conventions to standardize AI telemetry. The goal? Make AI agents traceable, debuggable, and less of a black box.
Choice cuts: The Best of What’s New
OTTL Playground for OpenTelemetry
OpenTelemetry Transformation Language (OTTL) is a powerful way to reshape telemetry data, but writing rules can feel like debugging regex—powerful, but painful.
Elastic recently introduced a new OTTL Playground that makes it easy to test and tweak transformations in real time, without breaking production.
Kasper, out.
