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

Episode 79 of The Site Reliability Podcast looks at how modern SRE teams are using observability tools to shrink mean time to detect — the gap between a system failure and the team knowing about it. Hosts Lucas and Luna break down why observability goes beyond traditional monitoring, using real-world examples like a major e-commerce platform that cut MTTD from 12 minutes to under 90 seconds by shi...

Lucas and Luna dive into the often-overlooked difference between Service Level Agreements (SLAs) and Service Level Objectives (SLOs) in site reliability engineering. They explore how SLAs are not just legal documents but critical tools for managing stakeholder expectations, using a real-world case from a major cloud provider. The episode explains the 99.9% vs 99.99% uptime debate, the cost implica...

Episode 77 of The Site Reliability Podcast dives into canary deployments: rolling out code changes gradually to a small subset of users before a full release. Lucas and Luna explain how companies like Netflix and Etsy use canary analysis to catch regressions early, using real traffic and metrics. They walk through the mechanics: routing a fraction of traffic, comparing key SLOs like latency and er...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into DORA metrics — the four key DevOps Research and Assessment measures that elite SRE teams use to quantify software delivery and operational performance. They break down each metric: deployment frequency, lead time for changes, mean time to restore (MTTR), and change failure rate. The hosts explain how Google's 2019 Accelerate...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the practical use of Service Level Objectives (SLOs) in site reliability engineering. They discuss how a major European bank reduced pager fatigue by 40% by shifting from alert-based monitoring to SLO-based error budgets. Lucas explains the difference between SLIs, SLOs, and SLAs, and why measuring user-facing latency is mor...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into how a blameless culture can actually improve incident response times and reduce recurrence. They explore a real case from a mid-size SaaS company that cut its mean time to resolution by 40 percent after adopting blameless postmortems. Lucas breaks down the psychological safety factors that make engineers more willing to shar...

In Episode 73 of The Site Reliability Podcast, Lucas and Luna explore how blameless postmortems transform incident response culture. Using examples from a major e-commerce platform's 2024 database outage, they break down the difference between blame and accountability, explain why 'human error' is a shallow root cause, and share how one team cut repeat incidents by 40% just by rewiring their post-...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how SRE teams apply fault tree analysis (FTA) from aerospace and nuclear engineering to reduce incident recurrence. Using a real 2025 outage at a major streaming platform where a cascading DNS failure took down services for 47 minutes, they break down the top-down logic of FTA, how it differs from postmortem 5 whys, and why te...

Episode 71 of The Site Reliability Podcast with Fexingo dives into how SRE teams are applying large language models and AI assistants to accelerate incident triage and root cause analysis. Lucas and Luna examine a real case from a mid-sized e-commerce platform: after a database connection pool exhaustion caused a 14-minute partial outage, the on-call engineer used a locally-run AI tool to correlat...

In this episode of The Site Reliability Podcast, Lucas and Luna dive into the practice of game days — structured simulations where SRE teams deliberately inject failures to test their incident response and on-call processes. They discuss a real example from a major streaming platform that runs quarterly game days to validate runbooks and reduce mean time to resolve from 45 minutes to under 15. The...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how error budgets help SRE teams make data-driven trade-offs between reliability and feature velocity. Using Google’s original framework and a real-world example from a major e-commerce platform, they explain how setting a 99.9% SLO with a 0.1% error budget per quarter creates explicit room for innovation without risking catas...

In Episode 68 of The Site Reliability Podcast, Lucas and Luna explore how SRE teams use infrastructure as code (IaC) to prevent configuration drift — the silent killer of production reliability. They break down a real incident at a mid-sized fintech company where a manual SSH change caused a partial outage, and how the team rebuilt their entire environment with Terraform and automated compliance c...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how SRE teams use incident response playbooks to standardize their reaction to common outages. They break down what makes a good playbook—specific, testable, and owned by a single team—using concrete examples like a Redis cluster failover and a database connection pool exhaustion. Lucas explains the difference between a playbo...

Site reliability teams spend huge effort on monitoring and alerting—but some of the worst outages start the moment a deployment goes live. In this episode, Lucas and Luna break down how readiness checks, or health probes, act as the first line of defense against bad code reaching production. Using the example of a major Kubernetes rollout gone wrong at a large e-commerce company, they explain the...

Episode 65 of The Site Reliability Podcast digs into a practical framework SRE teams use to tie infrastructure costs to specific services and teams. Lucas and Luna break down how cost attribution works, why it helps prioritize reliability investments, and a real example from a major streaming platform that saved millions by charging back observability costs to feature teams. Learn how to move from...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how SRE teams are using toil budgets to prioritize automation and reduce operational overhead. They dive into Google's original SRE definition of toil—manual, repetitive, non-value-added work—and explain how teams set toil budgets as a percentage of total engineering time, typically around 50 percent. The hosts discuss a real-...

In this episode of The Site Reliability Podcast, Lucas and Luna explore how SRE teams are shifting from reactive scaling to proactive capacity planning. They dive into the story of a major streaming service that averted a disastrous holiday outage by using predictive models based on historical traffic patterns and user growth data. The hosts break down the key metrics—like requests per second, mem...

In this episode of The Site Reliability Podcast, Lucas and Luna explore the art and science of capacity planning in SRE. They walk through a concrete case: how a major streaming platform used predictive modeling to avoid a holiday-season outage after underestimating user growth in a new market. Lucas breaks down the two main approaches — reactive vs. proactive planning — and explains why the best...

In Episode 61 of The Site Reliability Podcast with Fexingo, Lucas and Luna dive into chaos engineering—the disciplined practice of intentionally injecting failures into production systems to uncover weaknesses before they cause real outages. They explore the origins at Netflix and the emergence of tools like Chaos Monkey, Litmus, and Gremlin. The hosts discuss how SRE teams at companies like Amazo...

In Episode 60, Lucas and Luna dive into the most overlooked part of incident response: the postmortem action items that actually prevent the same outage from happening twice. They unpack a 2025 study from Google's SRE team that found 67% of postmortem action items are never completed, and explore why. Using concrete examples from a major AWS S3 outage and a Stripe payment-processing incident, they...

Episode 59 of The Site Reliability Podcast explores how SRE teams classify incidents by severity to decide how fast to respond and who to page. Lucas and Luna break down real-world classification frameworks — from SEV-1 (service down, all hands on deck) to SEV-4 (minor hiccup, fix in the next sprint). They discuss why vague severity definitions lead to alert fatigue and slow response times, and ho...

Episode 58 of The Site Reliability Podcast with Fexingo digs into post-incident reviews — not as blame sessions or compliance checkboxes, but as structured learning mechanisms. Lucas and Luna examine Google's seminal 2016 Titan key outage to illustrate how root cause analysis misses the point if teams don't ask 'why' five times. They discuss the difference between finding a 'root cause' and unders...

Lucas and Luna explore how SRE teams at companies like Spotify and Etsy use 'cost of delay' — a concept borrowed from product management — to quantify the business impact of reliability work. Lucas explains the math behind deferring a reliability project, using a real-world example: a payment-processing team deciding whether to fix a latency issue or build a new feature. Luna pushes back on the di...

Episode 56 of The Site Reliability Podcast explores how SRE teams at companies like Airbnb and Etsy use intelligent alert routing to slash incident noise by over 60 percent. Lucas and Luna break down the evolution from on-call pagers to modern event-driven routing, explain how machine learning models classify alerts by severity and team ownership, and discuss the trade-off between routing accuracy...

Episode 55 of The Site Reliability Podcast with Fexingo dives into incident cost analysis — a growing practice at companies like Google and Stripe where SRE teams assign a dollar value to every outage minute. Lucas and Luna break down the methodology: how to quantify direct revenue loss, reputational damage, and opportunity cost from incidents, and how that data helps teams justify automation spen...