Systems talk for performance-minded engineers.
PurePerformance brings performance engineering into the messy reality of modern software teams. Brian Wilson, Andy Grabner, and their guests connect AI, SRE, Kubernetes, open source, observability, and incident learning to the daily work of building systems that hold up.
PRESS PLAY
PurePerformance is a technically fluent show about the parts of software engineering that become visible when systems slow down, fail, scale, or change faster than teams can comfortably manage. Its center of gravity is performance, but recent episodes show how broad that word has become. AI tooling, code health, SRE practice, Kubernetes, resilience, cloud capacity, open source, and engineering education all sit inside the same operational frame. The conversations are interview-driven and practical. A typical episode starts with a concrete setup: a conference talk, a customer example, a guest’s work, or a small anecdote from the hosts. From there, the discussion widens into larger questions about process, architecture, feedback loops, and organizational learning. Brian Wilson and Andy Grabner bring a familiar back-and-forth, often mixing light teasing with serious follow-up questions. Guests are given room to explain, but the hosts regularly pause to clarify terms or translate an idea into plainer engineering language. That makes the show useful without making it overly polished. The recent AI episodes are especially focused on what changes when coding agents, memory systems, prompts, and AI-native workflows enter everyday development. The tone is curious but not breathless. The show asks how teams preserve fundamentals, review generated code, document intent, and keep future engineers from skipping the struggle that builds judgment. Its resilience episodes take a similarly grounded approach. Near misses, queues, capacity planning, incident learning, and postmortems are treated as clues about how complex systems really behave. Infrastructure conversations can get specific, including Kubernetes, multicluster batch dispatching, cloud instance constraints, and large-scale scientific workloads. Open source appears through career and company-culture stories rather than abstract advocacy. PurePerformance is best for listeners who already speak some software language and want sharper conversations about how engineering systems survive contact with production reality.
Made for: PurePerformance is for software engineers, SREs, performance engineers, platform teams, and technical leaders who want practical conversations about reliability, scale, AI tooling, and operational discipline.
What sets it apart: The show stands out by tying high-level shifts in software engineering to concrete production concerns: queues, code review, postmortems, cloud capacity, Kubernetes workloads, and the habits teams need when tools change quickly.
The brutal truth about digital performance engineering and operations. Andreas (aka Andi) Grabner and Brian Wilson are veterans of the digital performance world. Combined they have seen too many applications not scaling and performing up to expectations. With more rapid deployment models made possible through continuous delivery and a mentality shift sparked by DevOps they feel it’s time to share their stories. In each episode, they and their guests discuss different topics concerning performance, ranging from common performance problems for specific technology platforms to best practices in development, testing, deploying and monitoring software performance and user experience. Be prepared…
Based on 6 episodes we listened to · September 2026
PurePerformance connects observability, resilience and AI-era development to the operational decisions that determine how software behaves under pressure.
PurePerformance turns operational engineering into lively, guest-led conversations, with Brian Wilson and Andy Grabner moving between observability, site reliability, resilience, performance engineering and code prepared for agentic AI.
A one-hour queue timeout quietly loses orders; Andy Grabner and Brian Wilson make you check the defaults.
Press play if you want
Short reviews from the PlayNext desk, based on the episodes we processed.
A grounded look at AI-native work and the way interfaces and development habits may shift.
Chris Labrado joins Andy Grabner and Brian Wilson for a discussion about AI moving from helper to the thing shaping the interface. They work through AI-native artifacts, AI-native communication, and the practical question of whether email and PowerPoint still fit once more of…
A grounded conversation about code quality, AI readiness, and the value of effortful learning.
Pure Performance treats AI readiness as a matter of engineering discipline, not branding. Andy Grabner and Brian Wilson sit down with Adam Tornhill to talk through code hotspots, the code as a crime scene idea, and the habits that make a codebase easier for AI to work with.
A grounded look at AI agents, prompt engineering, memory, and the enterprise discipline needed for reproducible software.
PurePerformance widens its usual performance-engineering lens to ask what software work looks like when AI agents can build, communicate, and help shape go-to-market thinking.
A practical case for turning near misses into shared system knowledge before quiet failures become expensive.
PurePerformance returns to resilience with Adrian Hornsby, staying in the IT world while connecting chaos engineering to the discipline of learning from failure.
A practical SRE conversation about resiliency, capacity, and the tradeoffs behind keeping systems upright.
Pure Performance focuses on resiliency and reliability, with Alexandra Franz laying out the hardware constraints, cloud capacity, and tradeoffs that come into view when demand climbs.
A practical discussion of tracing, logging, and the tradeoffs behind simpler observability.
Pure Performance shifts into observability with Daniel Gomez Blanco, and it works best when it keeps traces, logs, and metrics in their own lanes instead of bundling them together.
A grounded look at performance work, open-source tradeoffs, and the conference circuit.
Pure Performance builds on Martin Spier's performance stories to tie together the hosts' seven-year run, episode 193, and the practical appeal of why Netflix opened up its tools.
Specialists walking through software's tradeoffs, one system at a time.
Founders and engineers on AI coding agents like Claude Code, and supply chain security.
Stack Overflow's Ryan Donovan interviews founders and engineers on AI agents, protocols, and dev tools.
Michael Kennedy and Brian Okken run down the week's Python tooling, typing, and packaging news.
Matt Stratton interviews DevOps practitioners on security, platform engineering, and trusting AI-written code.
Andy Grabner and Brian Wilson introduce a discussion on distributed tracing with Dynatrace engineers Christoph Neumueller and Thomas Rothschaedl.
Hosts Brian Wilson and Andy Grabner welcome Daniel Gomez Blanco, a principal engineer at Skyscanner and author of 'Practical Open Telemetry', to discuss adopting open observability.
Andy and Brian introduce the 10-year anniversary of their podcast and set up a discussion about MCPs with guest Dana Harrison.
Ricardo Rocha discusses how CERN processes 1 petabyte per second using Kubernetes for batch job dispatching.
Brian Wilson and Andy Grabner celebrate their 201st episode with guest Mark Tomlinson, discussing automation, AI, and CoPilot.
Brian Wilson and Andy Grabner introduce a guest, Liran Haimovitch, to discuss observability requirements for developers.
Martin Spier shares his 20-year journey in performance engineering, from early coding in BASIC to roles at Expedia and Netflix.
PurePerformance brings on Chris Labrado, a solutions architect at QVC group under the Home Shopping Network banner, who describes himself with 'life, liberty, and the pursuit of root cause.' He traces his path from bowling in 2014 and a 2019 YouTube channel called Bowler's Rant to building Dynaclaud, a browser tool that uses Dynatrace MCP and an agentic workflow for problem discovery and remediation.
Adam Thornhill returns to PurePerformance six and a half years after the first Code as a Crime Scene discussion, and he frames the bottleneck in software as understanding existing code rather than writing new code.
Andy Grabner frames Pure Performance around Wolfgang Heider's article "A Show Don't Spec.
Pure Performance brings back Adrian Hornsby, who says he left AWS after about four and a half years on the AWS Fault Injection Service team and started Resilium Labs to help organizations fix resilience programs.
Diana Todea shares her journey from being a beginner to becoming an open source contributor, offering advice on how to get started in open source projects.
Alexandra Franz, a Site Reliability Engineer at Dynatrace, explains how SRE work covers production, pre-production, monitoring, scaling, and on-call response across Austria, Detroit, Sydney, Texas, California, and other offices.