SREcon17 Europe/Middle East/Africa
2017-08-30..2017-09-01
Dublin, Ireland
| title | author | length |
|---|---|---|
| Care and Feeding of SRE | Narayan Desai | 35m |
Argues that SRE is an organizational strategy, not a bag of tactics, centered on three concrete goals: maintaining SLOs, managing operational load, and… | ||
| Diversity and Inclusion in SRE: A Postmortem | Niall Richard Murphy | 41m |
Minority engineers face higher attrition and documented bias: patches from women show 10% lower acceptance rate across 1.4M contributors, the industry bears… | ||
| Globalizing SRE in a Walkup Culture | Bill Lincoln, Matt Coleman | 22m |
Wayfair expands SRE from Boston into Berlin, managing ~5,000 production systems and >100 deployments per day, and builds a 24/7 NOC staffed in both regions… | ||
| Make Haste Slowly: Balancing SRE Diligence in Urgency... | Jason Hiltz-Laforge | 21m |
Shopify frames speed versus risk by quantifying a system’s cost of failure with a disaster matrix and an availability matrix that drives prioritization and… | ||
| Want to Solve Over-Monitoring... | Kishore Jalleda | 27m |
Company faces ~100,000 alerts per month; goal is less than two alerts per shift. Initiative named clean room negotiates alert budgets with each studio, denying… | ||
| SRE Your gRPC... | Gráinne Sheerin, Gabe Krabbe | 58m |
Shows concrete changes to a gRPC C++ stack: use Protocol Buffers for compatible APIs, add a translator backend, set per-RPC deadlines via flags, and propagate… | ||
| Profiling Node Applications | Sasha Goldshtein | 29m |
V8’s built-in profiler via --prof produces gigabytes of logs, only samples the main thread, and points to pbkdf2 using ~90% CPU in one example; perf attaches… | ||
| The Dangers of Being Overly-Paranoid | Ingrid Epure | 43m |
Intercom ships to production >100 times per day with 66 authors committing hundreds of times weekly, and postmortems show 70% of fixes are “add more data”… | ||
| Show Me the RIGHT Numbers! Are Our Users Happy? | Perry Statham | 21m |
Argues that reliability work must be outside-in SRE and measure actual user happiness not sysadmin metrics, using personas and scenarios to define SLIs/SLOs… | ||
| Standing On the Shoulders of Giants... | Emil Stolarsky | 23m |
Shopify embeds OpenResty with Nginx and LuaJIT to run Lua scripts at the edge, deployable in 10 second deploys, and uses a module called sorting hat with… | ||
| InStream: Large Scale Distribution... | Harsh Sharma | 22m |
InStream: Large Scale Distribution using BitTorrent, Python, Salt, and Kafka
Harsh Sharma, LinkedIn
Deploying application/services to all the servers across… | ||
| Use Load Testing to Build a Proper Mental Model of Your Service | John P. Looney | 31m |
Intercom builds a custom load generator, hooks the test cluster to production monitoring via terraform, and uncovers concrete failure modes: 35MB queries (1.5M… | ||
| Traffic Steering using Rum DNS @ LinkedIn | Abhijeet Rastogi | 40m |
Abhijeet Rastogi, LinkedIn
Do you serve customers across the world with varying network conditions? Do you struggle with automatically sending your users away… | ||
| Capturing and Analyzing Millions... | Karthik Appigatla, Basavaiah Thambara | 27m |
Capturing and Analyzing Millions of Queries without Any Overhead
Karthik Appigatla, Sr. Database Engineers, LinkedIn, and Basavaiah Thambara, LinkedIn
This… | ||
| OK Log: Distributed and Coördination-Free Logging | Peter Bourgon | 45m |
A distributed logging system that forgoes global coordination and instead writes ULID with millisecond prefix‑prefixed records into segment‑per‑connection… | ||
| How We Try to Make a Lion Bulletproof... | Janna Brummel, Robin van Zijll | 37m |
ING sets up SRE to fix repeated outages (2014 headlines) and raise 96.84% monthly availability toward regulator expectations. A small centralized SRE team… | ||
| From Firefighting to Proactive Work: ... | Alex Gerlic | 22m |
A major MySQL datastore hits a tipping point with a 66% rise in query volume, CPU often near 80% and frequent pages; a rapid firefight mobilizes ~25 engineers… | ||
| Incident Command at the Edge | Lisa Phillips | 36m |
Fastly describes a concrete, repeatable Incident Command protocol that coordinates SRE, network, security and support across 43 POPs and a 15+ Tbps edge… | ||
| Resiliency Testing with Toxiproxy | Jake Pittis | 18m |
Shopify built Toxiproxy to sit between app and service in dev and test and injects latency and black holes at the TCP layer so tests see realistic network… | ||
| Building a Culture of Reliability | Arup Chakrabarti | 34m |
PagerDuty describes concrete steps that make reliability measurable and actionable: pick customer-facing metrics (they use incidents resolved per hour for a… | ||
| Tech Leadership in SRE | Sean Rees | 31m |
Defines the tech lead as a role evaluated on team impact not personal output, and shows concrete practices: maintain broad context, draft a vision and roadmap… | ||
| Case Study: Lessons Learned from Our First Worldwide Outage | Yoav Cohen | 30m |
A dormant IncapRules double-quote parser bug triggers crashes on every proxy because the centralized Rev Lite leader repo replicates the malformed rule into… | ||
| When Trouble Comes to Town | Michael Gorven | 18m |
Seven practical steps guide response: quickly assess user impact with metrics and ask whether it affects 1% of requests or 100% of users; open an internal… | ||
| Building an SRE Capability Inside a Large Organization | Sriram Gollapalli | 41m |
A hardware-first company with 95% revenue from instruments moves from shipping CDs and 6–18 month releases to running SaaS and shipping twice a week by… | ||
| The Why, What, and How of Starting an SRE Engagement | Richard Clawson, Josh Gilliland | 23m |
Engagement with Azure IoT starts by taking over half the on-call rotation (eight developers initially thrown in) and uncovers missing runbooks, tribal… | ||
| Startup Systems Εngineer's Instruction Manual | Effie Mouzeli | 20m |
Jumping in as the sole systems engineer at a startup means owning the full stack: Salt Stack and Fabric for provisioning, Jenkins for automated deploys, a… | ||
| Cognitive Bias and On-Call | Niall Richard Murphy | 57m |
Surveys of ~200 SREs measure concrete biases: anchoring shifts time estimates when a 45s versus 30s prompt shows a sawtooth graph, ~30% fall for a Linda-style… | ||
| Reducing MTTR and False Escalations: Event... | Michael Kehoe | 30m |
LinkedIn builds an Event Correlation Engine that maps dependencies for 900 applications and 2,200 internal APIs into a call graph in a graph database, and runs… | ||
| The Never-Ending Story of Site Reliability | Kurt Andersen | 19m |
Argues that automation doesn't remove SRE roles and offers a measurable rubric: maps the Dreyfus skill levels to specific practices — incident response… | ||
| The History of How We Came to Be | Niall Richard Murphy | 21m |
Argues that engineering and operations split through three concrete forces: wartime staffing shifts that pushed women out of undifferentiated machine roles… | ||
| Hiring SREs May Be Literally Impossible | Chris Sinjakli | 21m |
Hiring external SREs is slow and lossy: GoCardless 2017 pipeline shows 160 applicants → 45 remote tests → 15 finalists → 2 offers accepted. The practical fix… | ||
| Gamifying Reliability Excellence—The Service Score Card | Daniel Lawrence | 23m |
A small web app called Service ScoreCard runs weighted checks (10–30 line Python plugins) against services by pulling data from Git, CI, config and monitoring… | ||
| Incident Management and Chatops at Shopify | Daniella Niyonkuru | 19m |
Shopify runs a dedicated IMOC rotation and a chatops bot called Spy (based on Lita) that binds a war room channel, auto-posts summaries, and integrates with… | ||
| Lightning Talks | Jaime Woo, Daniel Lawrence, Luiz Viana, Guy Fighel, Jason Koppe, Emil Stolarsky | 48m |
Six short reports deliver concrete results: a culture of communications outlines six rituals (RCAs published as newsletters, summits) that reduce duplicated… | ||
| Why Work with a Tech Writer? | Betsy Beyer | 12m |
Tech writing engagement turns a messy set of playbooks, runbooks, postmortems and QuickStart guides into maintainable, findable docs by choosing repositories… | ||
| Postmortem Action Items: Plan the Work and Work the Plan | John Lunney | 27m |
Postmortem practice centers on concrete, measurable follow through: classify action items as prevention, mitigation, detection, require at least one P0 or P1… | ||
| Building an On-Premise Kubernetes... | Daniel Turner | 19m |
Shopify migrates its Rails monolith and microservices from a homegrown system to on‑prem Kubernetes, building Etcd quorum (3/5/7), anycast master VIP for… | ||
| Distributed Systems, Like It or Not | Theo Schlossnagle | 27m |
Argument: distributed systems are now everywhere so failures matter because one node can affect 10 million browser nodes not just one user; timing and clocks… | ||
| Avoiding and Breaking Out of Capacity Prison | Jake Welch | 18m |
Describes a working capacity-management system used on Microsoft Azure Storage that automates cluster rebalancing, forecasting, and playbooked mitigations… | ||
| Service with an Angry Smile: Passive-Aggressive Behavior in SRE | Lauri Apple | 40m |
Argues that covert hostility in teams produces real SRE failures: an ignored alert that predicted an outage and a hoarded password causing a multi-hour outage… | ||
| The Cult(Ure) of Strength | Emily Gorcenski | 21m |
A culture that prizes visible sacrifice shows up as praise for being “brave” while managers ignore repeated misuse of names and pronouns — a formal complaint… | ||
| Run Less Software; Use Less Bits | Rich Archbold | 23m |
Intercom adopts the rule “run less software, use less bits” and keeps to a small set of standard technologies plus outsourced services using four outsourced… | ||
| Monitoring Cloudflare's Planet-Scale Edge Network | Matt Bostock | 31m |
Cloudflare runs Prometheus on each POP, federates selected metrics to four HA top-level servers, and reports concrete scale: 116 locations, 188 Prometheus… | ||
| Monitoring Design Principles | Theo Schlossnagle | 52m |
Reframes monitoring as measuring real system behavior not vague metrics, argues for SLOs that are assessable (calculate percentiles over fixed one to five… | ||
| And the CFO Wept: AWS Cost Control | Corey Quinn | 30m |
Tells how AWS bills balloon with concrete fixes: run a tagging audit to tag everything so finance can allocate costs; buy Reserved Instances to cut ~20% with… | ||
| Have You Tried Turning It off and Turning It on Again? | Tanya Reilly | 41m |
Argues that untested fallback plans and hidden service cycles, not abstract “backend” problems, make global restarts unrecoverable; cites real… | ||
| 100 Teams, 100 Ways to Fail | John Keiser, Ben Broderick Phillips | 40m |
Azure SRE describes engagements across 100+ teams, shows concrete fixes (a 40 lines of code chatbot to send group invites, service dashboards wired to live… | ||
| Persistent SRE Antipatterns: Pitfalls On the... | Jonah Horowitz, Blake Bisset | 52m |
Alerts target user-experience alerts not individual boxes: wake people only when a significant fraction of a cluster fails or velocity spikes, not when host… | ||
| nothing matches | ||