SREday London 2024
| title | author | length |
|---|---|---|
| Automated Incident Management | Ajuna Kyaruzi | 20m |
Meta uses retrieval plus large language modeling to rank root causes and finds it correct 42% of the time, Datadog adds AI-driven Slack incident summaries, and… | ||
| The History of DevOps in 30 minutes | Chris Cooney | 25m |
Covers a decade of operational shifts with concrete claims and engineering fixes: logs hit FAT32 4GB and inode limits, cloud cuts provisioning time… | ||
| Unleashing Potential Across Teams: Infrastructure as Code | Alayshia Knighten | 23m |
Demonstrates using Pulumi to provision a launch template and autoscaling resources from reusable components, with TypeScript components and stacks that map to… | ||
| Developers are all the same - right? | Joyce Lin | 20m |
Claims that developers are uniform get tested with concrete numbers and failures: Postman powers 30 million developers, Stack Overflow surveys 65,000… | ||
| Declarative Linux with the Common Operating System Interface | Justin Garrison | 24m |
Argues for replacing ad hoc server typing and SSH with a small HTTP/gRPC API surface that manages a purpose-built OS image for Kubernetes deployment… | ||
| The Evolution of Ops | Miko Pawlikowski | 14m |
Timeline of ops evolution claims that cloud native tooling explodes cognitive load with about 97 tools quoted to manage hundreds of microservices and that… | ||
| SREs are Driving the Evolution of Developer Self-Service | Sébastien Blanc | 24m |
Argues SREs cut developer cognitive load by turning platform actions into declarative APIs and automated pipelines, using Kubernetes as the base platform with… | ||
| Instant Clarity: Insights for SRE Excellence | Jerome Sadler | 24m |
Demoed application Neurologic ingests 10 GitHub repos (mix of public and private), builds up-to-date tech docs in ~2 minutes after ingestion deltas, and… | ||
| Improve application resilience with chaos engineering | Gunnar Grosch | 26m |
Claims that downtime costs enterprises up to $300,000 per hour (91% of surveyed firms) and roughly $10,000 per minute, then demonstrates resilience testing… | ||
| How to Destroy an SRE | Carly Richmond | 24m |
Argues that overloaded alerting, brittle tooling and poor processes actively destroy SREs rather than topics alone: migration to Helm packaging with explicit… | ||
| Beyond K8S - Maturing Your Platform Engineering Initiative | Nicki Watt | 27m |
Uses the CNCF Platform Maturity Model v1 to map concrete outcomes across investment, adoption, interfaces, operations and measurement and recommends plotting… | ||
| Platform Engineering: Evolution or a Fancy Rename? | Dotan Horovits | 25m |
Argues that platform engineering is a DevOps approach that turns reusable tooling into a product: Citrix centralizes certificate rotation and cuts integration… | ||
| Mystery of the disappearing S3 Keys | Simon Hanmer, Ross Walker | 15m |
Monitoring pipeline ingests ~1.5–2 million S3 object events per day and shows ~350 daily failures where code sees an object but S3 reports NoSuchKey… | ||
| What I Learned | Ian Miell | 25m |
A support org moves from ad hoc fixes to a documented, repeatable process by cataloguing 2,128 incidents over months, splitting teams into incident and problem… | ||
| From Hero to Zero: a guide to destroy your best employees | Antonio Cobo Cuenca | 28m |
Satire lists seven concrete ways managers break teams: micromanage with spy tooling (monitor typing, block autonomy), enforce arbitrary rules and… | ||
| The RunWhen Authors Community: Royalties For Runbooks | Kyle Forster | 22m |
Claims a platform that assembles community-written troubleshooting steps into goal‑oriented automation that runs in about three days instead of a year — 100x… | ||
| Next-Gen K8s Threat Hunting: GenAI & eBPF | Michele Dodic, Francesco Sbaraglia | 21m |
Claim: a working pipeline uses eBPF hooked to exec syscall to stream kernel events into a Python + OpenAI parsing pipeline that detects anomalous behavior… | ||
| From free-kicks to git commits | Steve Wade | 22m |
A career shifts from being scouted at four and suffering a career‑ending knee injury at 18 to building cloud native tooling: moves from AI-powered Kubernetes… | ||
| Crafting an SRE Roadmap for Enterprise Success | Jorge Luis Castro Toribio | 19m |
Worked with large enterprises to deliver a three‑milestone SRE roadmap: assessment‑goals‑culture where assessment measures observability, automation, testing… | ||
| Alerts don't suck, YOUR alerts suck | Leon Adato | 28m |
Argues that alerts work when they are actionable now, not vague FYIs, and shows concrete fixes: replace one-off CPU thresholds with a baseline plus threshold… | ||
| Use Karpenter to scale, optimize & upgrade K8s | Chakkree Tipsupa | 22m |
Karpenter replaces node pools and autoscaling groups and calls the EC2 API directly, cutting typical scale-up latency to about a minute in the demo (five… | ||
| Multi-environment deployment for K8s-based microservice arch | Carlo Ventrella | 18m |
Helm umbrella charts bundle dozens of microservices with per-customer values.yaml so product versions map to specific dependency versions and configs, enabling… | ||
| Virtual Kubernetes Clusters: A New Approach to Multi-tenancy | Hrittik Roy | 23m |
Claims that vcluster CLI creates full Kubernetes clusters inside a host cluster in under a minute, syncs pods and services back into renamed host namespaces… | ||
| When Infrastructure as Code Ends - Jump in and Create Some More | Harel Safra | 26m |
Claims that anyone can write a Terraform provider and shows how to do it with Terraform Plugin Framework and Go, using Delve for debugging, GPG-signed… | ||
| Architecture of a Fintech Startup | Aleksei Popov | 24m |
Stan builds a next‑gen capital platform and confronts heavy regulation and integration sprawl, with company onboarding split into 15 Steps and KYC/AML flows of… | ||
| Caching the uncacheable in Varnish | Thijs Feryn | 28m |
Explains how Varnish turns otherwise uncacheable responses into cached assets by using VCL to strip tracking cookies, create hash variants and override TTLs… | ||
| Mastering OpenTelemetry Collector Configuration | Steve Flanders | 29m |
Live demo shows an OpenTelemetry Collector YAML that actually runs: use the validate command to catch required fields before startup, start otelcol with a… | ||
| The House is on fire: mysteries of on-call K8s engineering | Alessandro Vozza | 22m |
KUspaces is pitched as namespace as a service built on vcluster and an Envoy Gateway and aims to hide cluster ops while exposing the Kubernetes API. Real… | ||
| The Subtle Art of Lying with Statistics | Dave McAllister | 27m |
Shows concrete ways numbers mislead: using arithmetic versus geometric means (reporting 1.5 hours vs 0.53 hours), hiding deviation (Death Valley mean 61°F… | ||
| Journey to Stable Infrastructures with Talos Linux & Cozystack | Andrei Kvapil | 25m |
Claims a cloud‑like, multi‑tenant Kubernetes on bare metal by combining Talos Linux, Cluster API and kamaji to run management control planes as Kubernetes… | ||
| Perses, the GitOps Native dashboard visualization tool | Erwin de Keijzer | 22m |
Perses claims to be a GitOps-native replacement for Grafana and is now in CNCF sandbox, but many goals remain unfinished. Demo uses Docker Compose with… | ||
| Unlocking key metrics and patterns using Grafana | Syed Usman Ahmad | 25m |
Shows how Grafana reads Google Sheets via the Google Sheets data source plugin using an API key or service account, demos building a table and time-series… | ||
| Improving the efficiency of more than 800 databases with o11y | Wilfried Roset | 22m |
OVHcloud builds an internal DB service that wires Postgres and MySQL logs and metrics into OpenSearch + Mimir + Grafana, parses slow queries by threshold, and… | ||
| DevOps Is Not Dead. Not Not by a Long Shot | Ricardo Castro | 17m |
Claims DevOps is alive after 15 years and remains the cultural foundation for faster, reliable software delivery; platform engineering is a complementary… | ||
| AI with Machine Learning for Observability in Netdata | Costa Tsaousis | 22m |
Netdata implements unsupervised anomaly detection at the edge, training a rolling model per metric every 3 hours on 6 hours of data, keeping 18 models and… | ||
| Vertically Integrated in Platform Engineering | Matt Simons | 17m |
Argues platform engineering should act like a vertically integrated product company that actually builds a developer platform (not the end application), tracks… | ||
| ChatGPT, LLMs, and LangChains - A Beginner's Guide | Marcin Szymaniuk | 21m |
Covers building end-to-end LLM systems with concrete tools: OpenAI chat API, LongChain chains that pipe prompts and outputs, connectors to PostgreSQL and… | ||
| New Approaches to Reduce Alert Noise AIOps | Birol Yildiz | 20m |
Platform uses offline open‑source vector embeddings stored in a vector database and a cosine similarity threshold to deduplicate alerts, removing… | ||
| Gen AI Powered Omni-functional Decision Insights for SREs | Mahesh Venkataraman | 19m |
Gen AI synthesizes telemetry and enterprise tool data to cut SRE cognitive load and pinpoint multi-domain causes of incidents, demonstrated on an e‑commerce… | ||
| Zero-Trust Security Strategy in Serverless App Deployment | Anthony Ekpechue | 20m |
Zero trust applied to a serverless e commerce wireframe claims to block breaches by making identity the perimeter and preempting compromise. The design uses a… | ||
| What does "high priority" mean? The secret to happy queues | Daniel Magliola | 27m |
Argues that queues must be organized by explicit latency guarantees (e.g., within 1 minute, within 10 minutes) rather than vague priority labels; implements… | ||
| Cloud Native Tech: The Building Blocks of Modern DB Software | Divine Akachukwu Odazie | 28m |
Traces the history from mainframes to cloud and containers and shows which Kubernetes pieces actually matter for databases: StatefulSets, PV/PVC and the CSI… | ||
| Kubernetes, AWS, ALBs, Terraform and no Helm! | Marcos Diez | 21m |
Argues that no Helm is safer: use Terraform (soon open tofu) to provision AWS ALBs, listeners, and empty Target Groups while the AWS controller adds pod IPs… | ||
| Config files vs. flags: story of pain | Oleg Fatkhiev | 21m |
Argues that preferencing command-line flags and environment variables over static files eliminates indirection and improves observability because the full… | ||
| Klustered: Kubernetes Debugging Live | David Flanagan | 35m |
Live debugging of Cluster 2 shows a deliberately sabotaged single-node Kubernetes cluster: missing kubeconfig, no CNI, CoreDNS pending and containerd running… | ||
| Oops, I deployed too hard | Matteo Bianchi | 22m |
A single bash script automates deploys and accidentally overwrites 419 with 420 during a Friday lunch, unlocking a laptop and triggering a customer outage… | ||
| Debezium Server - New CDC runtime on the horizon! | Ondrej Babec, Jiri Novotny | 26m |
Debezium Server appears as a standalone CDC runtime that replaces Kafka Connect’s embedded connectors with a three-stage architecture: source capture… | ||
| Using, and mis-using K8s Dynamic Admission Control | Richard Finlay Tweed | 20m |
Explains how Kubernetes dynamic admission control works, with mutating webhooks that rewrite images or inject default annotations and validating admission… | ||
| Essential Software Observability for Every Role: Look No Further | Piyush Verma | 32m |
Argues that observability is about detecting unexpected change, not nicer dashboards, citing 18 deploys per day and the ladder from basic monitoring → faster… | ||
| nothing matches | ||