Today is another big day for @dash0hq. We're welcoming @PolarSignalsIO to the team. Polar Signals has built the leading continuous profiler for cloud-native workloads, including GPU and CUDA support, and created Parca, the open source continuous profiling project. Together, we're bringing continuous profiling natively into Dash0. That means profiling won't live in a separate tool; it’ll sit alongside traces, metrics and logs in the same OpenTelemetry-native platform, unifying the data and giving engineers a much deeper understanding of what's happening inside their applications. We've shared more about why we made this move in the comments.

You’ve got an SLO. Your error budget is disappearing. Now what? Liat Hoffman takes just over two minutes to show you how they work in practice, from setting a reliability target and tracking burn rate to seeing which deployment might be behind the change. Watch the breakdown below, and if you want to go deeper, Liat’s full SLO webinar from earlier this week is now on YouTube: eu1.hubs.ly/H0yM4wS0

114 new teammates in the first half of this year. Two acquired teams joining Dash0. Growing that quickly changes a lot more than the number of people on Slack. How you hire, how you onboard, and how you keep teams connected all start looking different at that pace. That’s some of the context behind Dash0 joining Batch #5 of Scale-up.NRW, alongside other growing companies from across North Rhine-Westphalia. Paloma Baltazar, our VP Operations, joined the Full Circle Night to kick things off and meet some of the founders, mentors, and investors involved in the program. We’re looking forward to learning from the other companies in the batch, comparing notes on the challenges that come with growth, and seeing what we can bring back to Dash0 along the way. #ScaleupNRW #Batch5 #NRW Ministerium für Wirtschaft, Industrie, Klimaschutz und Energie des Landes NRW Start2 Group digihub.li Photo by paint the town studios

Mirko Novakovic was invited into the Nasdaq studio to talk about #Dash0, #AIagents, and where #observability is heading next. It was another special moment for Dash0, just months after our $110M Series B put us on the Nasdaq screen in Times Square. Watch the full Nasdaq interview here: eu1.hubs.ly/H0yvrbm0

Most teams adopt @opentelemetry to escape vendor lock-in. That undersells why it became the standard behind modern observability and AI. Join Julia Furst Morgado, Principal Developer Relations Engineer at Dash0, for an OTel 101 on how it works and why every kind of AI depends on structured telemetry. 📆 October 6th, 11am ET / 5pm CEST 🎁 Get a coffee voucher for joining, and a chance to win a bigger prize if you stay until the end. Save your spot here: eu1.hubs.ly/H0yyHGP0

Hotservability is the live #Observability talk where the observability questions get harder as the wings get hotter and hotter. This week, we’re live at the 7th @TheLeadDev Stockholm Meetup. Dash0 is sponsoring the evening at Umain on October 1st. Julia Granström opens with the future of software engineering. Marcus Hammarberg continues by talking about navigating uncertainty in the #AIera. Then Michele Mancioppi and Sazzad Khan take over, and it’s Hotservability time! Somewhere between increasingly difficult questions and unreasonably hot sauce, we'll find out whether Michele and Sazzad can withstand the heat. Doors open at 17:30, with food and drinks throughout the evening. And yes, you can try the hot sauces as well! Spots are limited and RSVP is required. If you're in Stockholm, grab your spot here: eu1.hubs.ly/H0yFKzM0

Code RED Newsletter #38 is out - Kasper's fortnightly read on what's moving across observability, OpenTelemetry, and AI. This issue gets into a theme that keeps coming up: understanding what systems - and increasingly agents - actually did, rather than relying on what they say they did. - Uber stopped asking people to maintain dependency maps and started working them out from production failures. At 0.01% sampling, ten failed traces could take two days to collect, so they use metrics. - Atlassian moved metrics for ~100,000 hosts across 14 regions to the OpenTelemetry Collector. 4.8 billion data points a minute go in, and about 220 million come out. - The Kubernetes attributes processor reached v1.0.0 and is now stable for logs, metrics, and traces, while the Collector releases and the Collector as a whole are not at v1.0.0 yet. - Henrik Rexed benchmarked OTel-Arrow and found that the CPU savings didn't go away. They moved it to an upstream Collector. Also in there: Satya Nadella on containment and auditability, Mirko Novakovic on agents that fail while every dashboard stays green, SLOs and AI SDLC Insights in Dash0, Julia Furst Morgado on why OpenTelemetry alone doesn't end vendor lock-in, and a rogue goose at SRE Day London. Full issue here: eu1.hubs.ly/H0yC9n_0

The red jumpsuits are a familiar sight at WeAreDevelopers in Berlin by now. This week, they’re making their debut in San José. 🇺🇸 Around 30 of us made the trip for the first WeAreDevelopers World Congress in North America. The past 24 hours have mostly been flights, setup, and slowly turning one corner of the McEnery Convention Center a bit more red. Now the doors are open, Booth #743 is up and running, and the first demos are underway. If you’re here, come by and see Dash0 in action.

If dashboards are built for humans, what do AI agents need? It’s not more dashboards. That’s one of the things Mirko Novakovic gets into in his conversation with Yossi Garmazi and Jason Woolf from Boardroom Club. Humans are good at looking at graphs and spotting patterns. Agents need something else. They need the right tools, the right context, and interfaces like MCP. As Mirko puts it, the number of human users is staying more or less flat, while the number of agents is exploding. They also talk about OpenTelemetry, telemetry growth, Agent0, AI inference costs, and building in the AI era. Watch the full episode here: eu1.hubs.ly/H0yx_Hq0

Your dashboards are green, nothing crashed, and nobody got paged. Meanwhile, your AI agent could be doing exactly the wrong thing. That’s the problem Mirko Novakovic digs into in his latest post, starting with Satya Nadella’s description of AI agents as a new kind of insider risk and his call for aggressive monitoring of agent activity. Traditional observability is good at telling you when a system fails. Agents introduce a different problem. They can stay operationally healthy while doing something you absolutely didn’t want them to do. If an agent is going to act on its own, you need to be able to reconstruct what it did. And knowing what it did is only part of it. You also need to know whether what it did was actually good. That’s where observability, evals, security, and DevOps start to come together. Read Mirko’s take here: eu1.hubs.ly/H0ywx8v0

AI writes more of your code every sprint. The question that never gets easier is whether the next one is safe to ship. Join Liat Hoffman, Product Manager at Dash0, for a walkthrough on building an #SLO on a real service, plus open Q&A. 📆 September 29th, 11am ET / 5pm CEST 🎁 Get a coffee voucher for joining, and a chance to win a bigger prize if you stay until the end. Save your spot here: eu1.hubs.ly/H0ylJLv0

Service Level Objectives are live in Dash0, and today we’re on Product Hunt. Here’s the problem we kept running into: the service is up, traffic is flowing, and someone asks, “Is it reliable enough?” Ask three engineers, and you’ll probably get three different answers. An SLO gives everyone the same reference point: a target, a rolling window, and an error budget. Once that budget starts disappearing, the question is how fast. A bad deployment can burn through a big chunk of it in an afternoon. A slower issue might chip away at it for days or weeks. Dash0 tracks the burn rate and puts deployment events right on the SLO view, so you can see what changed. Setting one up is simple. Define good and total events, set the target, and Dash0 handles the recording rules underneath. You can do it in the UI, with Agent0, or through OpenSLO, the API, and Terraform. 25 SLOs per organization, included. You can find us on Product Hunt today: eu1.hubs.ly/H0yr7PF0

Yesterday, we wrapped up LDX3 New York. 🗽🍎 Julissa Mackey, Matthew Galvin, Jordan Steinberg, Sean P, Michael Ehrlich, Andrew Contreras, Mara Natale, and AJ Audino held down the booth at the Javits Center. If the red jumpsuits caught your eye and stopped you for a chat, it was great talking to you. On to the next one!

Reliability, as a number. Set a target. Watch the error budget. See how fast you're burning it. One or two PromQL expressions to define an SLO. We build the recording rules underneath. Deployment events show which release broke it. 25 SLOs included per org. Dash0 SLOs are now GA: eu1.hubs.ly/H0ylJJ-0 Join our SLO webinar on September 29: eu1.hubs.ly/H0ylJLv0

London brought the energy last week! 🇬🇧 A few of the #Dash0 team spent the week in the UK, seeing plenty of familiar faces and meeting a lot of new ones too. We kicked off with a breakfast for a small group of engineering leaders, getting into what it actually takes to run AI and agentic workflows reliably in production. Then it was London Tech Week and a run of local meetups. We spent the rest of the week moving between sessions, coffees, and good conversations. Thanks to everyone who grabbed a coffee or stuck around to chat. Our travel calendar is getting pretty full for the next few months, so chances are we’ll see some of you again soon.

There is no secret sauce. That’s what Mirko told James Pringle on the #RidingUnicorns podcast. Building a company is just bricks, stacked one at a time, and the hard part was never finding a rare brick. It was having the patience to let the stack compound. Here's the part that gets wild. Almost no code at Dash0 gets written by hand anymore. Engineers work in natural language and spend their time reviewing what AI produces instead of typing it line by line. Some developers spend twenty thousand dollars a month on AI tools. No cap. We just built something internal to track the spend against what it returns, and that tracker turned out good enough to become a product of its own. Mirko also opens up about the exact moment Dash0 hit real momentum, and why he pushed harder than he had in his last two companies combined. This one is worth the full listen. Spotify: eu1.hubs.ly/H0ydrmd0 Apple: eu1.hubs.ly/H0ydrv30

Two conferences, eight time zones apart. Julia Furst Morgado was at #KCD San Francisco. Kasper Borg Nissen was at #ContainerDays in Hamburg. By Friday they had the same conversation about six times. Almost none of it was about models. Engineers wanted to know who watches an agent once it is loose in production, and where organizational knowledge lives so a coding agent can use it instead of guessing. The pager is not impressed by your model. The model is the part you can buy. The floor underneath it is the part you have to build. That is the throughline in Code RED Newsletter # 37, Kasper's fortnightly read on the ecosystem. - monday.com standardized alerts across 600+ services before they let an agent triage on top of it. - #Uber's coding agents run on a 24 million node context graph. Without it, one query took 20 minutes. With it, 38 seconds. - Mirko Novakovic asks #Cribl CEO Clint Sharp why he thinks the AI SRE category is dead on arrival, on Code RED Podcast # 49. - Michele Mancioppi runs the #OpenTelemetry Collector on #NixOS and gets a validation default flipped upstream. Also in there, Kasper on Dash0 joining the #AgenticAIFoundation, Alex Palcuie at #Anthropic sitting with why Claude still cannot fix itself, and how #Figma keeps different kinds of agent memory deliberately separate. Full issue here, link below.👇 eu1.hubs.ly/H0yg3Sh0

By lunchtime at most conferences, everyone on the team has seen a different talk and you end up comparing notes on what everyone missed. #SignalsConference is set up differently. One track, one room, around 250 industry experts following the same sessions for two days. That’s a big part of why we’re excited to sponsor Signals this year. Some of the #Dash0 team are here in Berlin, and later today Mirko Novakovic is joining Niall Murphy, Rodrigue Schäfer, Hannes Lenke, and Alejandro Saucedo for the closing panel on Reliability in the Age of AI. We’re at Kulturbrauerei today and tomorrow, so if you’re at Signals, come by and say hi!

Hamburg last week. Montreal this week. Booth number two of September is underway, with Matt, Troy, Thor, Reece, Tommy, and Brennan for RustConf. Booth T6 is where you’ll find us over the next couple of days. Come by, show us what’s giving you trouble, and we’ll show you how we’d approach it in #Dash0. See you there. 👋

Dash0 is now a Silver member of the Agentic AI Foundation. Open standards are at the core of how we build Dash0: OpenTelemetry, PromQL, and Perses. We don’t want proprietary formats to be the reason anyone stays with us. We think the same should be true for agents. The agentic ecosystem is moving fast, and there’s still a lot to figure out around how agent behavior should be described consistently. Which tools did it call? What was it trying to accomplish? What happened across an entire run? These are questions we’re already running into ourselves while building and instrumenting agentic systems. So it makes sense for us to be part of the conversations where those standards are being shaped. Kasper Borg Nissen has already been involved since joining the first AAIF Ambassador cohort in June. Dash0 joining as a member takes that involvement a step further. Read more from Kasper on why we’re joining and where we want to contribute: eu1.hubs.ly/H0y7Y2T0

We’ve had more than one Leadership Dinner this year that was supposed to end at 21:00 and was still going at 22:30. That’s probably the best explanation for why we keep doing them. A small group of engineering and technology leaders, good food, and the kind of conversations you rarely get to have at a busy conference. Whether it’s OpenTelemetry, observability costs, AI in production, platform engineering, or whatever is proving difficult that week, there’s room to actually get into it. We’ve already started September with dinners in Santa Monica and Hamburg, with London, Paris, and Amsterdam still to come. If you’re in one of those regions and interested in joining us, or want us to bring one closer to you, drop us a message. We’ll see what we can do. 😄