The Site Reliability Podcast with Fexingo: SRE, Uptime, and Production Engineering

Most SRE teams talk about incident response and automation, but fewer talk about the human side of on-call: how to pay people fairly for the disruption. Lucas and Luna dig into a 2025 survey of 500 SREs that found 62% feel on-call pay does not match the cognitive load. They compare models — flat stipend versus per-incident pay — and discuss how companies like Honeycomb and PagerDuty structure thei...

Site reliability engineering has a well-known failure mode: your pager goes off at 2 AM for a minor blip, or worse, you don't get paged until a full-blown outage has already hit users. This episode explains SLO burn rate alerts — a concept that Google's SRE team refined in their 2016 book and which is now baked into tools like Google Cloud Monitoring, Datadog, and Grafana. Lucas and Luna walk thro...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the growing importance of the Software Bill of Materials (SBOM) for securing software supply chains. They use the 2024 XZ Utils backdoor as a concrete case study to explain how a single maintainer burnout led to a critical vulnerability that an SBOM could have caught earlier. Lucas breaks down what an SBOM is, how it works w...

In Episode 51 of The Site Reliability Podcast, Lucas and Luna explore how SRE teams use feature flags—not just for canary releases, but as a core tool to decouple deployment from release, reduce blast radius, and enable instant rollback without redeploying. They walk through a real incident at a major streaming company where a misconfigured flag caused a 47-minute partial outage, and how the team...

Lucas and Luna explore how production stress testing goes beyond standard load testing to simulate realistic user behavior, with a deep dive into how a major streaming platform used session replay and gradual ramp-up to validate infrastructure before a global event. They unpack why stress testing must replicate authentication flows, API call patterns, and edge case traffic shapes — not just raw re...

Lucas and Luna explore how site reliability engineering teams use game days — structured, simulated incident exercises — to prepare for real outages. They break down the approach used by a major fintech company that runs quarterly game days for its entire on-call rotation, with concrete scenarios like a simulated database failover and a DNS misconfiguration. The episode covers how game days differ...

In episode 48 of The Site Reliability Podcast with Fexingo, Lucas and Luna dive into error budgets — the SRE concept that turns reliability into a business decision rather than a purely technical one. They break down how Google originally defined error budgets via the Service Level Indicator (SLI) / Service Level Objective (SLO) / error budget framework, then explore how teams at companies like Sh...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the often-overlooked first step of SRE practice: defining Service Level Indicators (SLIs). They explore how vague uptime percentages fail to capture user experience and walk through a concrete example from a major streaming platform that shifted from a 'five nines' target to a more granular SLI based on video start latency....

Episode 46 of The Site Reliability Podcast with Fexingo dives into how SRE teams are applying cognitive load theory to reduce burnout and improve incident response. Lucas and Luna explore the concept of 'cognitive load' — the mental effort required to operate complex systems — and how teams at companies like Google and Netflix use techniques like toil reduction, documentation, and team topologies...

Episode 45 of The Site Reliability Podcast digs into observability—how modern SRE teams go beyond monitoring to discover the 'unknown unknowns' that cause the worst outages. Lucas and Luna break down the difference between watching known metrics (CPU, memory) and exploring unknown failure modes with structured events and high-cardinality data. They walk through a real example: a major e-commerce p...

In this episode of The Site Reliability Podcast, hosts Lucas and Luna explore how SRE teams at major tech companies build and maintain dependency graphs to predict cascading failures before they happen. Using concrete examples from cloud infrastructure and microservices architectures, they explain how graph-based service maps help teams identify single points of failure, model blast radius, and pr...

Episode 43 of The Site Reliability Podcast. Lucas and Luna explore how SRE teams are adopting 'toil budgets' — a concept inspired by error budgets — to cap the amount of manual, repetitive work engineers do each sprint. They break down Google's internal definition of toil (hands-on work with no enduring value), how a toil budget works alongside an error budget, and a concrete case from a mid-sized...

This episode explores how Site Reliability Engineering teams use Service Level Objectives (SLOs) not just as a quarterly dashboard metric, but as a real-time decision-making tool that shapes pager rotations, deployment gating, and incident prioritization. Lucas walks through how Shopify's SRE team used a 99.95% availability SLO to flag a critical degradations before it became a full outage in 2025...

Lucas and Luna dive into canary deployments: the practice of routing a small percentage of production traffic to a new version before rolling it out broadly. Lucas explains why Netflix's 'canary clusters' and Etsy's 'feature flipping' approach revolutionized how SRE teams think about release risk, and contrasts it with the old all-at-once deploys that caused major incidents. They discuss specific...

In episode 40 of The Site Reliability Podcast, Lucas and Luna dive into chaos engineering — the practice of intentionally breaking systems to find weaknesses before real incidents strike. They explore how Netflix pioneered the approach with Chaos Monkey, the lessons SRE teams can learn from controlled failure experiments, and how to start small with simple game days that simulate a database partit...

Episode 39 of The Site Reliability Podcast with Fexingo dives into capacity planning as a proactive SRE practice. Lucas and Luna explore how teams at companies like Google and Netflix use trend analysis, load testing, and headroom budgeting to avoid capacity-related outages. They discuss a real-world case from 2025 where a major streaming service averted a Super Bowl crash by scaling capacity week...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how SRE teams use immutable infrastructure to eliminate configuration drift and improve reliability. They dive into a real case from Google's Borg paper, explaining how replacing mutable servers with golden images reduces incident rates and recovery times. The hosts break down the trade-offs with mutable servers, the role of i...

In this episode of The Site Reliability Podcast with Fexingo, Lucas and Luna explore how SRE teams are using auto-remediation to automatically resolve incidents without human intervention. They break down the anatomy of an auto-remediation pipeline — from monitoring alerts to automated runbook execution — using real-world examples like a major streaming service that reduced pager fatigue by 40 per...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the incident command system (ICS) model that large-scale SRE teams borrow from emergency services to manage complex outages. They walk through a real example: a major payment processing incident at a fintech company where a database migration triggered a cascading failure affecting three million users. Lucas explains the fou...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how blameless postmortems go beyond simple incident analysis to drive real systemic improvements. Using the example of a major payment processor incident in early 2026, they break down the anatomy of an effective blameless postmortem: separating human error from system design flaws, writing actionable recommendations, and trac...

In this episode of The Site Reliability Podcast, Lucas and Luna dig into the one incident-documentation practice most teams get wrong: the postmortem. Most postmortems are filed and forgotten. Lucas walks through how Google's SRE team shifted from blame-free to action-oriented postmortems, using a concrete example from their own 2017 Gmail outage. He breaks down the difference between a cause and...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the practical side of runbook automation — moving beyond static documentation to executable, automated responses. They explore how companies like Google and Netflix use runbook automation to reduce mean time to repair by up to 60%, and discuss the common pitfalls: over-automation, stale runbooks, and the tension between spee...

In this episode of The Site Reliability Podcast with Fexingo, Lucas and Luna dive into the often-overlooked intersection of site reliability engineering and cloud cost optimization. They explore how SRE teams at companies like Uber and Airbnb use techniques such as right-sizing instances, leveraging spot instances, and implementing autoscaling policies to reduce infrastructure spend without sacrif...

When a massive traffic spike hits, every millisecond of latency can cost thousands of dollars. In this episode, Lucas and Luna explore load shedding — the SRE technique of intentionally dropping non-critical requests to keep core systems running. They walk through how Google SREs used load shedding during the 2020 YouTube outage, how Stripe applies graceful degradation during payment surges, and w...

Feature flags are a powerful tool for SREs, but they come with their own operational risks. In this episode, Lucas and Luna explore how companies like Etsy, Netflix, and LaunchDarkly use feature flags to decouple deployment from release, enabling canary rollouts, instant kill switches, and safer experimentation. They break down the difference between boolean flags, multivariate flags, and experime...