Dash0

When every client has a different stack, cloud provider, and way of working, observability can quickly turn into something you rebuild over and over again. For viadee Unternehmensberatung AG, every team had its own dashboards and alerts, every new project came with another round of setup, and when something broke in production, logs and metrics could only take them so far. With Dash0, they’ve been able to change that. viadee now has a setup they can reuse across client environments. They can get meaningful traces within one or two days, move between logs, metrics, and traces without piecing everything together manually, and use Agent0 to spend less time figuring out where to look in the first place. As Maximilian Bastin, IT Consultant and Cloud Engineer at viadee, explains: "We went from spending time managing observability infrastructure to actually using observability to build better software." Read the full viadee story: https://lnkd.in/ebkrvE6m

The latest G2 results brought some great recognition for Dash0, from Best Support and Most Implementable to Users Most Likely To Recommend. Thanks to everyone who took the time to share their feedback. 💙

Polar Signals is joining Dash0. It started with Mirko Novakovic and Frederic Branczyk seated next to each other at a Berlin dinner, talking storage engines. Not the version you'd put in a deck. Frederic tells his side on Code RED Podcast #48: profiling was always powerful, it just needed an expert to read it. That's the lead in Code RED Newsletter #36, Kasper Borg Nissen's fortnightly read on the ecosystem. The rest is quieter work, mostly about making what you already collect honest enough to act on. -> Google's A2A protocol is now an AAIF project, sitting next to MCP. Different jobs, same neutral governance, no fork. OpenTracing and OpenCensus did this and became OpenTelemetry. -> Arize donated the code behind OpenInference to OpenTelemetry's GenAI SIG. Not a merger, and gen_ai.* stays put. -> OpenTelemetry entity events keep the stream of what changed and when. Overwrite the state and you lose both. -> Severin Neumann on database telemetry: the slowest query isn't the one worth fixing. Multiply duration by call frequency and the list reorders. The team is out twice next week. Julia Furst Morgado talks OTTL with Reese Lee at KCD San Francisco Bay Area on 1 September. Kasper takes the OpenTelemetry maturity model to ContainerDays Hamburg, 2-4 September, on a historic cargo ship. Say hi 👋 Also in there: Andrew Mallaband on "production understanding,” Ayooluwa Isaiah watching Claude Code's own token spend, Kasper on observability's Rails  on charts, and a switch that makes Agent0 ask before it writes. 👉 Full issue: https://lnkd.in/g7yUXD9W

Our Ecosystem Engineer showed up on Vietnamese national television a few weeks ago, and we're still thinking about it. Mauricio (Salaboy) Salatino was at KCD Vietnam, addressing the country's tech industry on AI sovereignty. Telcos, banks and government institutions are racing to get their data centers ready for AI workloads, and a camera crew was in the room to catch it. The segment aired on national TV. We've seen Dash0 on the Nasdaq billboard in Times Square. Now we can add Vietnamese national news to the list. Check out Mauricio's segment! 🇻🇳📺

60 million spans per second. That is what Zalando's observability platform has to absorb during Cyber Week. Heinrich Hartmann joined Mirko Novakovic to talk through what it actually takes to run observability at that scale: 3,500 microservices, 250 Kubernetes clusters, and the sampling and cost control decisions that make it survivable on call. One of the more concrete moments in the conversation is Heinrich explaining why Zalando standardized on OpenTracing and then OpenTelemetry across the entire estate rather than letting teams pick their own path. The answer involves coordination costs that most orgs only feel after it is too late to fix cheaply. Full conversation: https://lnkd.in/ge_NyHai

In 2019, Bryan Liles challenged the Kubernetes community to find its "Rails moment." In many ways, platform engineering became that moment. Absorb the complexity centrally and give developers good defaults, golden paths, and escape hatches when they need them. Observability is still looking for its equivalent. We have powerful tools and increasingly mature standards like OpenTelemetry. But developers are still too often expected to assemble the experience themselves. Instrumentation, dashboards, alerts, metadata, and correlation across signals. The next step is treating observability as a platform product. Build the defaults into the platform. Automate the undifferentiated instrumentation. Give developers observability by design, while leaving them responsible for the business context only they can provide. And as AI agents increasingly reason over telemetry, getting that foundation right matters even more. Kasper Borg Nissen explores what observability's "Rails moment" could look like, and how we get there. Read the full piece at https: https://lnkd.in/gkkssPNe

Zero-touch instrumentation means getting OTel running without writing a single line of instrumentation code added manually to your applications. On August 26th, Michele Mancioppi (Chief Architect and Head of Open Source at Dash0) joins “What's new in OTel,” on behalf of the OpenTelemetry Packaging SIG live session, to talk through exactly how that works and where it is heading. He will cover the new Linux auto-instrumentation system packages and why the community sees packaging as a core adoption mechanism, not an afterthought. If you work on platform engineering or manage OTel rollouts, this one is worth your time. https://lnkd.in/gZ2RYcRU

Expedia's new AI agent STAR has no function calling, no memory, and no autonomous tool use. That's not a limitation. It's the design. Meanwhile Instacart's agent Blueberry jumped from mid 60 percent diagnostic accuracy to the high 90s. The upgrade wasn't a better model. It was fourteen years of incident history. Two teams made opposite bets on autonomy and landed in the same place. Kasper Borg Nissen is back with Code RED Newsletter 35, catching up on everything the ecosystem did while Dash0 was busy announcing Polar Signals. 🔗 https://lnkd.in/eqA44H6K

The Dash0 x Polar Signals deal was born at a dinner table in Berlin. "We've been circling similar problems from different directions for a while. Now we get to solve them together." Read the story from Mirko Novakovic and Frederic Branczyk: https://lnkd.in/eddiram7

Frederic Branczyk has watched customer after customer discover the same thing: profiling data that's been sitting there the whole time, untouched, while their cloud bill climbs. One customer cut AWS spend by 35% just by acting on it. He joined Mirko Novakovic to talk about why that data goes unused and the story behind Polar Signals joining Dash0. Episode out now on Code RED. Link in the comments.