SREcon18 Europe/Middle East/Africa
2018-08-29..2018-08-31
Dusseldorf, Germany
who it's for
site reliability engineers, engineering managers, platform engineers, on-call engineers, devops engineers, privacy and compliance officers, incident responders
topics
GDPR compliance and DPIAs; standardized postmortem corpora; canary deployments at scale; Kubernetes deploys and migrations; serverless chaos engineering; real world SLOs and SLIs; observability for emerging infrastructure; multi-region availability strategies; migrating to stateless cloud services; capacity planning and triage math; documentation and run books best practices; debuggable Go servers
| title | author | length |
|---|---|---|
| Design (Failures) Case Study | Theo Schlossnagle, Heinrich Hartmann | 41m |
Circonus describes building a custom C message queue and a bespoke columnar store (IronDB) to replace RabbitMQ and Postgres, achieving ~1M msgs/sec per… | ||
| SRE Theory vs. Practice: A Song of Ice and TireFire | Corey Quinn, John P. Looney | 31m |
A postmortem of polished conference rhetoric vs messy reality: a touted “perfect” SRE setup actually hides chaos engineering used as blame insurance… | ||
| Data Protection Update and Tales from the Introduction of the GDPR | Simon McGarr | 26m |
GDPR is live but enforcement ramps up: the CJEU faces Privacy Shield likely to be struck down which would immediately cut off transatlantic transfers unless… | ||
| What Makes a Good SRE: Findings from the SRE Survey | Dawn Parzych | 36m |
Survey of 416 responses analyzed with pivot tables and open text coding finds over 200 unique job titles and most SREs have ~5+ years experience. Top technical… | ||
| Sustainability Starts Early: Creating a Great Ops Internship | Fatema Boxwala | 27m |
Intern programs focus on concrete deliverables: realistic projects, good docs, frequent feedback and mentorship. Examples include 12–16 weeks internships that… | ||
| The Silver Lining Consortium: Post-Mortems for the Rest of Us | Niall Murphy | 34m |
Argues for a consortium that defines a standardized post-mortem schema and runs a centralized, searchable corpus of public cloud post-mortems with tooling to… | ||
| Migrations under Production Load: How to Switch... | Vilde Opsal | 24m |
Four-phase checklist: justify the change, set up parallel writes, compare results, optimize rollout, then wait and clean up. Two real projects show the… | ||
| The 7 Deadly Sins of Documentation | Chastity Blackwell | 46m |
Covers seven concrete failure modes and fixes backed by a GitHub 2017 stat that 93% of projects find docs insufficient and by Yelp’s experience creating a… | ||
| Y2K and Other Disappointing Disasters: Risk Reduction and Harm Mitigation | Heidi Waterhouse | 40m |
Y2K required thousands of engineers to edit mainframe date logic and reboot servers that sometimes never came back, and minor July outages still occurred… | ||
| Availability, Latency, and Cost: Withstanding Regional Outages | Aaron Blohowiak | 39m |
Netflix moves from an island model to a 3-region (2+1) N+1 model, cutting failover time from ~45 minutes to 7 minutes with Project Nimble and running… | ||
| SRE for Mobile Applications | Samuel Littley | 38m |
Google Search app SRE work focuses on client-side monitoring, progressive rollouts and remote feature flags to avoid rollbacks; real tactics include UI tests… | ||
| Your System Has Recovered from an Incident, but Have Your Developers? | Jaime Woo | 37m |
A 40-person straw poll finds nearly half of engineers feel stressed after incidents; over 50% report worse sleep and concentration, ~66% worse mood, ~50% less… | ||
| Against On-Call: A Polemic | Niall Murphy | 49m |
Argument: humans perform poorly under on-call stress—baseline error rate ~0.5% and errors double under time pressure (Microsoft chess study), 26% vs 19% report… | ||
| Impact of Network Automation | Roman Romanyak | 41m |
Small network team moves internal routing from OSPF to BGP with a reusable Ansible roles workflow (Jinja2 templates, Netconf via Python libs) and automates… | ||
| Migrating Your Old Server Products to Be Stateless Cloud Services | Kurt Scherer, Craig Knott | 35m |
Migrates Jira Data Center VMs into AWS using Packer, Ansible, Terraform to produce AMIs, Terraform modules to instantiate VPCs/ALBs/ASGs with RDS and EFS, and… | ||
| Lightning Talks | Daisy Galvan, Will Nowak, Kyle Lexmond, Dominic Green, Stephan Erb, Effie Mouzeli, Vladimir Dobriakov, Jorge Cardoso, Jason Gwartz | 46m |
Nine eight-minute case studies that report concrete fixes and results: a messy discovery service refactors to an object-oriented OS-class mapper + factory… | ||
| Dealing with Dark Debt: Lessons Learnt at | Vanessa Yiu | 30m |
Describes concrete practices at Goldman Sachs: a 25‑year SEC DB ecosystem with its own language slang, fortnightly releases, automatic pruning of unused code… | ||
| Halt and Don’t Catch Fire | Effie Mouzeli | 19m |
Argues that new small ops teams can’t erase legacy debt but can slow its creation by identifying five stages of grief for technical debt, building a visible… | ||
| Applying the Principles of Chaos to Serverless | Yan Cui | 43m |
Applies chaos engineering to AWS Lambda by instrumenting clients and configs rather than killing servers, using X‑Ray and SSM Parameter Store to injectableAll… | ||
| Know Your Kubernetes Deploys | Felix Glaser | 28m |
Shopify enforces deploy-time signatures on container images so only artifacts that pass precomputed checks can reach Kubernetes: Kritis (binary authorization)… | ||
| Not Invented Here Syndrome and Dark Debt: The PagerDuty Story | Aish Raj Dahal | 29m |
PagerDuty builds a Cassandra‑backed distributed queue called WorkQueue in 2011 to get high availability, durability and synchronous replication across vendors… | ||
| Building a Debuggable Go Server | Keeley Erhardt | 25m |
Improbable extracts common production plumbing into a reusable Go base server called radical that sits under gRPC/HTTP services, wires protobuf-generated… | ||
| Building a Fellowship Program to Mentor and Grow Your SRE Team | Tom Spiegelman | 23m |
DigitalOcean runs a two-week fellowship that embeds one developer with senior infra engineers, using handshake pairing, one-on-ones, exit interviews and a… | ||
| Story of Seeking Sustainable SRE | Björn Rabenstein | 29m |
SoundCloud has ~100 million users but only 100 engineers, so microservices outnumber people and Google-style SRE fails when copied directly. The company builds… | ||
| How We Un-Scattered Our DNS Setup and Unlocked New Automation Options | Dan Lüdtke | 22m |
Company owns over a hundred domains spread across five registrars and 30+ name servers, causing slow lookups and rollout limits. Teams consolidate registrars… | ||
| Kernel Upgrades at | Pradeep Nayak, Udupi Kadbet | 27m |
Facebook automates kernel upgrades for tens of thousands of database hosts using a fleet of Python state-machine jobs that call kernelctl install kernel X… | ||
| Managing Misfortune for Best Results | Kieran Barry | 39m |
Describes a weekly outage training game called wheel of misfortune with a “victim” (next on-call), a dungeon master who drives the scenario, a scribe, and… | ||
| Clearing the Way for SRE in the Enterprise | Damon Edwards | 45m |
Enterprises fail at SRE because org scar tissue — silos, queues, excessive toil, low trust — prevent the three SRE essentials: service level objectives with… | ||
| The Math behind Project Scheduling, Bug Tracking, and Triage | Avery Pennarun | 44m |
Presents concrete rules and simulations that actually predict schedules: an ad‑hoc simulator SimSWE models spikes when features change late, showing a team of… | ||
| Ethics in Computing | Theo Schlossnagle | 40m |
Computing grows from a 25-year industry to something that touches billions without domain-specific ethical enforcement, so engineers build defeat devices like… | ||
| Canarying Well: Lessons Learned from Canarying Large Populations | Štěpán Davidovič | 40m |
Argues that canarying faces a concrete tradeoff between fast rollout, small blast radius, and metric coverage and gives rules to balance it: pick a handful of… | ||
| Real World SLOs and SLIs: A Deep Dive | Matthew Flaming, Elisa Binette | 37m |
Defines SLOs from the outside in by mapping system boundaries to concrete capabilities, then measuring SLIs (eg. 99.9% uptime target for platform ingest and… | ||
| I’m SRE and You Can Too!—A Fine Manual... | Blake Bisset, Jonah Horowitz | 52m |
Recommendation: build a small, funded SRE core that proves value with tooling and training, not org fiat. Key moves: implement CI/CD and a deploy/monitor… | ||
| Lessons Learned—Data Driven Hiring 3 Years Later | Chris Stankaitis | 35m |
Pythian shifts hiring from trivia about tools to measurable potential: baseline technical checks plus structured evaluation of five to seven nontechnical… | ||
| SRE Team Lifecycles | Stephen Thorne | 47m |
Argues SRE is a lifecycle built on three concrete rules: set measurable SLOs and attach an error budget policy, give SREs bounded operational time so they have… | ||
| Capacity Planning in Four Parts: Telling the Future without a Crystal Ball | Evan Smith | 35m |
Four clear steps: pick an SLO and translate business risk into an error budget, find driver metrics by aligning SLIs with resource inflection points, turn… | ||
| The Nth Region Project: An Open Retrospective | Andrew Bloomgarden | 37m |
New Relic rebuilds an EU region in IBM Frankfurt to run ~500 services from ~50 teams, converting services to containers, standardizing CoreOS, Vault for… | ||
| This IS NOT Fine: Putting Out (Code) Fires | Emily Freeman | 27m |
Argues for adopting firefighting-style incident management: formalize an incident commander role, a comm chief who records timestamps and schedules the post… | ||
| What Medicine Can Teach Us about Being On-Call | Daniel Turner | 23m |
Hospitals cut harm with concrete tactics SRE teams can copy: a 5-step checklist collapses a 27-step process and drops infection rates from 11% to 0% in one… | ||
| Tradeoffs in Resiliency: Managing the Burden of Data Recoverability | Kristina Bennett | 31m |
Five years of Google data-integrity work shows concrete tradeoffs: redundancy is not durability because bad writes replicate; transaction logs often only… | ||
| Scalable Coding—Find the Error | Igor Ebner de Carvalho | 19m |
Shows concrete failures and fixes: regex backtracking can explode latency (a hand-rolled split is ~8× faster than the regex example), throwing an exception… | ||
| Delete This: Decommissioning Servers at Scale | Anirudh Ra | 19m |
Facebook runs decommission at scale across 12→14 sites and tens of thousands of servers and replaces a months‑long, spreadsheeted process with a UI plus a job… | ||
| Observability for Emerging Infra: What Got You Here Won't Get You There | Charity Majors | 46m |
Modern infra breaks dashboards and metrics; observability is the ability to ask new questions of raw, wide events not metrics and get answers. Build structured… | ||
| Deploying SRE Training Best Practices to Production... | Jennifer Petoff | 35m |
Google SRE education team builds a hands‑on photo-upload training service deployed to production without real users, runs monthly nuclear production… | ||
| Keep Building Fresh: Shopify's Journey to Kubernetes | Niko Kurtti | 46m |
Shopify moves a monolith and hundreds of apps onto Kubernetes over about a year, proving a platform approach that hits ~80% of use cases and rolls changes… | ||
| The Myth of Cloud Agnosticism | Corey Quinn | 38m |
Argues that true cloud agnosticism rarely exists: a survey of ~1,000 readers shows CloudFormation 43% / Terraform 41% and most companies run one primary cloud… | ||
| SRE for Good: Engineering Intersections between Operations and Social Activism | Liz Fong Jones, Emily Gorcenski | 36m |
Engineers are told to prevent harm by acting early: catch biased models (one government classifier was changed so it if true return detain), stop platforms… | ||
| Can I Tell You a Secret? I See Dead Systems | Avishai Ish Shalom | 15m |
Modern systems carry 40+ year old cruft that still shapes behavior: sending an interrupt to the keyboard controller on every boot to switch CPU mode (the 8042… | ||
| Junior Engineers Are Features, Not Bugs | Kate Taggart | 33m |
Claims that hiring juniors forces fixes in three areas—architecture, process, tooling—are backed by examples: a team with one third code‑school grads runs ~85%… | ||
| nothing matches | ||