#44 – Tracing at Planet Scale: How Zalando Monitors 60 Million Spans Per Second with Heinrich Hartmann

Heinrich Hartmann joins Dash0’s Mirko Novakovic to share how Zalando runs their observability at an extraordinary scale. Heinrich explains why Zalando standardized on OpenTracing and OpenTelemetry across 3,500 microservices and 250 Kubernetes clusters, what it takes to process up to 60 million spans per second during Cyber Week, and how the team approaches sampling, cost control, and adaptive telemetry. They also discuss the future of observability in the age of AI and Heinrich’s new conference, SignalsConf Berlin, which will focus on reliability and AI.

Buoyant CEO and creator of Linkerd William Morgan joins Dash0’s Mirko Novakovic to explore what happens when you instrument traffic at the service mesh layer. Drawing on his experience at Twitter, William explains how Linkerd gives platform teams HTTP and gRPC-level visibility without touching application code, why retry budgets help prevent cascade failures, and how workload identity creates a stronger foundation for security and observability. They also discuss MCP traffic, AI agents in production, and how observability needs to be re-optimized for LLMs.

Jad Naous, founder and CEO of Grepr, joins Dash0’s Mirko Novakovic to tackle one of observability’s biggest problems: too much data and not enough signal. Drawing on his experience at AppDynamics and Apache Druid, Jad explains why current architectures are fundamentally broken, and how Grepr compresses, aggregates and extracts signal directly in the stream, reducing data volume by up to 90% before it ever hits tools like Datadog. They dive into real-time pattern detection across logs and traces, why rule-based pipelines fall short, and how high-signal data is the foundation for any future AI SRE system.

Abby Bangser, founding principal engineer at Syntasso and co-author of “Platform as a Product,” joins Kasper Borg Nissen to unpack why most internal platforms struggle, and what it means to treat them like products. They explore producer vs. consumer dynamics, why golden paths often fail, how to measure platform success, and the shift toward internal “platform marketplaces” that scale beyond a single team. The conversation also covers observability for vs. of the platform, and why AI is putting even more pressure on platforms to remove bottlenecks and make access to resources feel effortless.Links mentioned in this episode:Platform as a Product report: https://www.syntasso.io/platform-as-a-product-oreilly-report

Better Stack co-founder and CEO Juraj Masar joins Dash0’s Mirko Novakovic to challenge the fundamentals of modern observability, from cloud lock-in and pricing models to how platforms should be built in the age of AI and how we market them. They discuss why observability costs are fundamentally broken, how Better Stack combines cloud and ‘bare metal,’ and why small teams can outperform large engineering orgs. The conversation also explores eBPF as the new default for instrumentation, the shift toward AI SRE, and how Better Stack has paired a developer-first product with unconventional marketing, from generous free tiers to SEO-driven status pages to a massive YouTube presence.

incident.io co-founder and CEO Stephen Whitworth joins Dash0’s Mirko Novakovic to explain why paging someone is only the start of an incident, not a holistic solution. They break down how incident.io supports the full incident lifecycle (coordination, comms, timelines, and follow-ups), why incident response is a “multiplayer game” across engineers, support, and leadership, and how AI is starting to reshape triage by pulling context from telemetry, past incidents, and customer signals. The episode closes with a practical look at what it will take to safely move from AI-assisted response to AI-driven auto-fixes that minimize the rollout ‘blast radius.’

Mauricio Salatino, open source and ecosystem engineer at Diagrid and author of Platform Engineering on Kubernetes, joins guest host Kasper Borg Nissen to break down why Kubernetes is a foundation, not just a standalone platform. They discuss ecosystem-driven platform design, reducing developer cognitive load, bringing feedback from production back into the ‘inner loop’ of development, and how generative AI is reshaping platform APIs and tooling. They explore whether Kubernetes is still waiting for its “Rails moment” — an opinionated, developer-first layer that makes building on it dramatically simpler.Links mentioned in the episode:Platform Engineering on Kubernetes by Mauricio Salatino (Manning):https://www.manning.com/books/platform-engineering-on-kubernetesThe Evolution of Platforms: Gen.AI Edition (Blog Post by Mauricio Salatino):https://www.salaboy.com/2025/11/18/the-evolution-of-platforms-genai-edition/

Checo, CEO and founder of Ottermon AI, joins Dash0’s Mirko Novakovic to argue that modern observability is rife with noise, reactivity and human bottlenecks. Drawing on years of frontline SRE and product experience, Checo explains why most telemetry is wasted, how signal distillation and “fingerprinting” can surface real risk earlier, and why observability must shift from dashboards and alerts to prescriptive, prevention-first intelligence.

Marylia Gutierrez, principal software engineer at Grafana Labs and an OpenTelemetry maintainer, joins Code RED guest host Kasper Borg Nissen, Dash0’s principal developer relations engineer, for a deep dive into how OpenTelemetry really works. They unpack how contributors get started, SIGs, why non-code contributions like documentation, localization, and governance matter just as much as PRs, and what it takes to grow from first-time contributor to maintainer. Links mentioned in the episode:How to Contribute to OpenTelemetry, by Marylia Gutierrez: https://opentelemetry.io/blog/2025/contribute-to-otel/OpenTelemetry community resources: https://opentelemetry.io/docs/contributing/

NOFire AI founder and CEO Spiros Economakis joins Dash0’s Mirko Novakovic to discuss why traditional observability and post-incident RCA are no longer enough in an AI-accelerated engineering world. Drawing from real production failures and years of SRE leadership, Spiros explains how causal AI and agentic workflows can predict failures before code ever reaches production. The conversation also explores why the future of observability is moving from dashboards and alerts toward understanding, reasoning and proactive decision-making.