OpenTelemetry for Balatro

I should start this blog post with a witty joke, because that’s one of the key ingredients that make Balatro such a beloved game. However, I can’t come up with one, so here’s a different intro: I am fascinated by the possibility that any software I use could, in theory, emit logs, metrics and traces through OpenTelemetry. And Balatro is software that I use a lot. I’ve spent hundreds of hours with it, the best 2 × 10 euros I’ve ever spent1. So creating OpenTelemetry for Balatro was a side project I wanted to try out for a long time! It turned out to be one of those projects where I cared more about the outcome than about writing the code myself, so it was a perfect candidate to ask an LLM to build for me! ...

2026-11-12 · Severin Neumann · Blog

Trace Through: End-to-End Observability from Claude Code to Your Tools

Claude Code session traces can leave the tools they invoke disconnected from the agent’s trace. In this episode, I show how to propagate trace context with CLAUDE_CODE_PROPAGATE_TRACEPARENT and connect agent activity, Docker builds, and custom tools into one continuous trace. The walkthrough covers Claude Code configuration, OpenTelemetry settings for Docker BuildKit and Buildx, and instrumenting your own tools using context propagation through environment variables. ▶ Watch on YouTube

2026-09-22 · Severin Neumann · Video · via Bronto

How to Expand the Observable Universe: 4 Techniques Every Developer Should Know | CloudNative Munich

In this talk from the Kubernetes, Cloud Native and Platform Engineering Meetup Munich, I explore observability from first principles and explain how to add OpenTelemetry to applications and their dependencies. The talk covers trace context, APIs and SDKs, OTLP, and semantic conventions before comparing four automatic instrumentation techniques: monkey patching, bytecode instrumentation, compile-time instrumentation, and eBPF. It also explores native instrumentation, instrumentation libraries, and practical steps for contributing to the OpenTelemetry ecosystem. ...

2026-09-09 · Severin Neumann · Video · via Bronto

Metrics are like Champagne

I always thought that metrics were like water. As long as something looks like one and is used like one, it is a metric. However, I was wrong. Metrics are more like champagne: only what comes from the Champagne region may be sold under the name. Everything else is sparkling wine. Likewise, metrics are only metrics if they were produced in the Vallée des Métriques. Everything that does not originate from there is just a sparkling roll-up. ...

2026-08-10 · Severin Neumann · Blog

Reliability from Manual to Autopilot

“AI SRE” is everywhere right now — but what does it actually mean to automate reliability, and how far can we really take it? In this talk from Cloud Native Karlsruhe, I break down the hype around AI SRE by comparing it to a discipline that has already gone through decades of automation: autonomous driving. Just like cars combine deterministic systems (anti-lock brakes, gearboxes) with AI-driven ones (adaptive cruise control, lane keeping), reliability engineering needs the same nuanced mix — not a single LLM doing everything. ...

2026-08-05 · Severin Neumann · Video · via Bronto

Observability is an infinite game, not a war.

If you’re interested in observability, it is very likely that you read Mat Duggan’s Why Clickhouse is Winning the Observability Wars or one of the follow ups and responses to it, for example Charity Majors’. Both are fantastic articles, and I recommend that you spend the time reading them front to back, give them your attention! And by that I also want to get one thing out of the way, that the title here might suggest: I do not disagree with them, I actually agree wholeheartedly with them on why ClickHouse is the current dominant backend in Observability. And they deserve that spot, as I know first hand from working with their product and knowing some of the great people working on it. ...

2026-07-09 · Severin Neumann · Blog

Adjusting load generators for realistic traffic simulation

Throughout my jobs in the observability space, I created or contributed to various demo and sample applications, which often follow the same premise: there is a “normal state” in which the application is running, and with a trigger, it moves into a “deviated state”. For example, there is the placeOrder transaction on a webshop that performs just fine, and orders and money are flowing into our hypothetical e-commerce company. However, with the click of a button (or a CLI command), an issue is injected into the application, and the placeOrder transaction stops working as expected. Orders go down, money stops flowing, hypothetical customers get angry! ...

2026-02-23 · Severin Neumann · Blog

Are metrics the bestrics?

“What’s your favorite telemetry signal” has been the last question of the Humans of OpenTelemetry series for the last few years. At KubeCon EU 2024 my answer to this was “profiling, because I think this is really closing a big gap that was missing in observability”. But, today, I found out that my answer has changed, and I am leaning more toward what Vijay Samuel gave as an answer: “I feel metrics are the most powerful signal!” ...

2026-02-10 · Severin Neumann · Blog

Can you get Observability without Telemetry?

People always say there are no stupid questions, and then you read the title of this post and you’re not so sure anymore. You start to doubt my sanity, or at least suspect that I’m a troll. However, as it is with most apparently stupid questions, there is something to learn from the answer if you explore it. To spare you from reading the rest of this, the short answer is “Yes, but…” and the long answer is more of a theoretical observation with some linguistic subtleties. So if you’re not interested in that, you can leave and do something fun, otherwise don’t say I didn’t warn you! ...

2025-12-18 · Severin Neumann · Blog

Splitting out a monolith into multiple services in OpenTelemetry

I did an experiment on splitting out a monolithic application into multiple “virtual services” in OpenTelemetry to have modules visualized independently on service maps. I am not sure if this is a good idea and something you should replicate in practice, since it might violate some best practices. However, I wanted to see how I can do it. Since (as far as I know) all otel backends are only able to provide such a map/graph visualization using service.name from the resource attributes, I tried out what happens if I create one TracerProvider per module with module-specific service.* attributes. ...

2025-11-25 · Severin Neumann · Blog