SREday 2023 London
| title | author | length |
|---|---|---|
| Futuristic Luxury Incident Response | Mandi Walls | 31m |
Argues that incident response improves with concrete changes: automation as first responder (20–30% of incidents resolve within five minutes and shouldn't wake… | ||
| Supercharging Observability with Feature Flagging | Alayshia Knighten | 33m |
Combines feature flags with Honeycomb observability to track a real rollout of service maps, instrumenting traces with a feature key so every user journey… | ||
| Fake It 'Til You Make It: Get the Most out of Incident Simulations | Ashley Sawatsky | 30m |
Run timeboxed (e.g., 90 minutes), realistic mock incidents that are designed to fail on purpose so gaps surface: use a facilitator, responders, silent… | ||
| The state of SRE in 2023 | Miko Pawlikowski | 16m |
Claims that SRE still lacks a clear definition and suffers from branding chaos, backed by an informal poll showing even splits between DevOps, SRE, production… | ||
| The Big One | Ajuna Kyaruzi | 26m |
Datadog suffers a global outage when an unattended systemd security update on Ubuntu restarts thousands of Kubernetes nodes and deletes their psyllium CNI… | ||
| Failing to Autoscale? | Erwin Daria | 31m |
Claims that machine learning right-sizes Kubernetes pod CPU/memory requests to stop the cluster autoscaler from adding costly nodes, reporting about 50%… | ||
| Five Challenges in Real-time Stream Processing and Five Solutions | Fawaz Ghali | 30m |
Hazelcast-based pipeline demo connects Redpanda (Kafka) topics, generates orders with Python, joins streams in an in‑memory IMAP and runs SQL-like queries to… | ||
| Lessons learned after more than 20 years working in IT | Antonio Cobo Cuenca | 28m |
Twenty years in IT distills to concrete practices: a first job with 32 new hires and no senior wastes time and money when juniors don’t ask questions, so build… | ||
| Cracking Cholera’s Code: Victorian Insights for Today’s Technologist | Simon Copsey | 26m |
Victorian London faces a cholera crisis with ~100,000 dead by 1854 and twenty years lost to the wrong theory; John Snow's cholera map and counting deaths… | ||
| Cloud Fitness Engineering - Maximizing Business Value from Cloud | Mahesh Venkataraman | 29m |
Accenture finds only about 37% of firms realize full cloud value and the talk argues over 30% of cloud projects miss business outcomes; Cloud Fitness… | ||
| Lessons learned from 7 years of running developer platforms | Michael Coté | 33m |
Seven lessons from seven years distilled into concrete practices: product-manage the platform using surveys and cycle metrics, prefer starting with one app… | ||
| Overcoming SRE Anti-Pattern Roadblocks: Rebranding the Ops Team | Ricardo Castro | 29m |
At FanDuel (Flutter/Blip) an SRE team retools incident management, observability and SLOs instead of just renaming ops: they own a SaaS observability stack… | ||
| Platform Engineering meets Service Mesh: an SRE love story | Alessandro Vozza | 33m |
Argues that adding Istio and Envoy gives platform teams concrete wins: Istio installs a control plane and sidecars to capture every request, delivers Envoy… | ||
| Navigating GraphQL Vulnerabilities: SRE / DevSecOps | Antoine Carossio, Tristan Kalos | 15m |
Escape scans 1,500+ production GraphQL endpoints using a feedback-driven API exploration engine and 500+ compute hours, finding 46,000+ security issues with… | ||
| On-premise data centers do not need to be legacy | Fabio Alessandro Locati | 30m |
Argues that private data centers match cloud economics by adopting precise practices: charge apps with fine-grained usage to reveal true cost, standardize a… | ||
| My Journey to SRE | Termitope Faro | 33m |
An engineer explains how SRE combines product development and operations by actually building developer-facing tools: a service catalog that returns deployment… | ||
| Deploying applications to Kubernetes using Helm | Andrew Pruski | 33m |
Helm packages Kubernetes YAML into versioned charts and the demo builds, packages and publishes a chart to a GitHub chart repo, then installs SQL Server images… | ||
| Building a Viral Open Source AI Chatbot: Journey from Concept to Reality | Matt Carey | 24m |
An open source chatbot called Quivr goes viral in about two months, claiming 14k stars and 6.5k active users, built as a document-embedding RAG assistant with… | ||
| Smart Contract-driven AI: A New Era of Decentralization | Lucas Roitman | 33m |
Claims a workflow to upload TensorFlow or PyTorch model to blockchain and run inference across a node network coordinated by a smart contract coordinates node… | ||
| Slaying the SLAs: Mastering Effective Communication | Krishna Pomar | 19m |
Built an internal app called Roadrunner using Python FastAPI, Celery, Redis, MongoDB and Jinja2 to manage ~3 billion daily messages and meet a 90‑day archival… | ||
| From Docker'fail to Dockerfile.. a true story! | Mohammed Aboullaite, Djalal Elbaz | 38m |
Session compiles 10 concrete Dockerfails (9 categories plus a bonus) and shows fixes with runnable tooling and a GitOps pipeline demo (repo linked). Specifics… | ||
| Distributed Caching in Real-Time Payment Systems | Aman Sardana, Bhargav Nachegari | 54m |
Built an active‑active distributed cache that supports thousands of TPS with sub‑second responses and five nines availability, switching from server to… | ||
| The efficient way of Autoscaling Your Workload in Kubernetes | Henrik Rexed | 40m |
Baseline cost is about $9 for the demo app. Using k6 load tests, OpenCost and Istio/Envoy metrics, iterative tuning of requests and limits drops cost to about… | ||
| DevOps 2.0 - Bigger, Badder with More Automation | Yishai Beeri | 31m |
Data shows pull requests sit idle about 50% of the time and a third idle for 80% of their lifespan, creating long context switches and risk. A measurable… | ||
| Synthetic Monitoring and E2E Testing: 2 Sides of the Same Coin | Carly Richmond | 26m |
Demonstrates using Playwright tests stored in a repo as journeys as code, pushed by GitHub Actions to Elastic Synthetics via API keys so the same scripts run… | ||
| Chaos Synergy: A Game Theory Approach to Chaos Engineering | Madhu Kumar Reddy | 24m |
Claims an AI model Chaos Synergy that combines game theory (uses Nash equilibrium) with ML to pick multi‑component chaos experiments, auto‑generate experiment… | ||
| Stop configuring infrastructure, start coding it! | Martin Sakowski, Julian Lang | 40m |
Argument: replace hand-editing JSON/YAML with real code using the open-source AWS Cloud Development Kit (CDK). CDK compiles TypeScript/Python/Java/C#/JS/Go… | ||
| Know your data: The stats behind your alerts | Dave McAllister | 41m |
Choice of central measure changes alerting: an arithmetic mean reads 13.33 req/s while a geometric mean compresses that to ~6.52 and a harmonic mean sits… | ||
| Feature Flags with OpenFeature: Faster, More Collaborative Development | David Hirsch | 43m |
OpenFeature is an open specification that lets applications swap feature-flag backends with a few lines of code — the demo shows a provider change of about 30… | ||
| Reliability Engineering: What we learnt so far | Ruggero Tonelli, Forlidar Macias | 28m |
Netquest runs seven engineers responsible for 90 production platforms across six Kubernetes clusters and ~100–150 autoscaling groups, supports ~70 product… | ||
| Statuscake vs Kibana alerting | Virginia Diana Todea | 25m |
Elastic replaces external uptime checks with in-house Synthetic monitoring that triggers Kibana alerting, automating monitor deployment from a GitHub repo… | ||
| Data Radar Maps: A Snapshot of Your Organizational Data Framework | Joy Chatterjee | 25m |
Data Radar Maps present a five‑segment scoring framework (organization, taxonomy, tools & processes, data fluency, data utilization) that turns 50‑question… | ||
| Gain a deeper understanding of your incidents: continuous profiling | Christian Simon | 41m |
Continuous profiling with Pyroscope and pprof pinpoints real root causes: an incident shows 800k goroutines chewing gigabytes of memory due to per-line logging… | ||
| Centralise legacy auth at the ingress gateway | Andrew Kirkpatrick | 27m |
Built a Flask auth service running as a standalone Kubernetes deployment behind Emissary Ingress that normalizes five legacy auth types (Basic, bearer, API… | ||
| Meet Kairos, an o-s project building the immutable Kubernetes edge | Nicolas Vermande | 37m |
Kairos converts any Linux container image into a bootable ISO and an immutable OS with active/passive partitions, TPM disk encryption, and a recovery… | ||
| We Stopped Thanos from Snapping $100K from Infra | Shubham Srivastava, Deepak Kumar | 27m |
A chain of concrete misconfigurations drives the story: a missing S3 VPC endpoint forces all Thanos sidecar uploads over the public internet, exploding egress… | ||
| We need to talk about SRE burnout | Ludovic Farine | 18m |
Incidents remain common and costly—25% of incidents cost over $1M and 30% of outages last more than 24 hours—and burnout comes when meaningful work is replaced… | ||
| Data Reliability Engineering: Secret Of Our Success | Akshay Karle, Carmen Mardiros | 31m |
Built a harmonized reporting product for a large publisher using data contracts, CI/runtime validation, and production monitoring; input violations most often… | ||
| Revolutionizing DevEx: Accelerating Software Development | Pavlos Kleanthous | 16m |
A UK client with 50-engineer teams adopts Infrastructure as Code and CI/CD to give each developer ephemeral per-PR tenant environments running on a common AWS… | ||
| Automation Best Practices for SRE and Security | Sarjeel Yusuf | 33m |
Datadog builds a workflow automation product that ties alerts to actions (Slack, EventBridge, AWS) and learns four concrete practices: design reusable modular… | ||
| Culture Trumps YAML | Brian Murphy | 27m |
Argues that fixing culture yields bigger wins than better YAML, showing a personal shift from coding to coaching and process work; implements an on-call rota… | ||
| Mastering Incident Communications | Kat Gaines | 26m |
Support teams act as the primary customer liaisons and must craft public messages using tools like PagerDuty, Jira, Slack, status pages and ticketing… | ||
| Kubernetes, Decisions | Kobi Biton | 36m |
Seven concrete decisions for running Kubernetes at scale: choose tenancy model (trusted, trusted soft isolation, or untrusted), decide cluster count (customers… | ||
| Unwiring High Cardinality | Piyush Verma | 56m |
Shows why a single payment timing metric easily multiplies into 100K combinations and how that scales to millions of active time series, causing queries that… | ||
| Observability Visualization in the Age of OpenTelemetry | Devrim Demiroz | 43m |
OpenTelemetry actually unlocks concrete visuals not vague promises: engineers extract sequence diagrams from traces to validate OIDC/SAML flows across 100+… | ||
| nothing matches | ||