SREcon22 Americas
2022-03-14..2022-03-16
-- online --
who it's for
site reliability engineers, observability engineers, platform engineers, engineering managers, incident responders, engineering recruiters, sre mentors and trainers
topics
self-destructing feature flags; eBPF-powered observability; OpenTelemetry tracing for bare metal; metrics-driven prioritization; modeling alert quality; serverless functions for observability; operational debt analysis; securing the software delivery chain; exemplar tracing practices; principled performance analytics
| title | author | length |
|---|---|---|
| The 'Success' in SRE Is Silent | Casey Rosenthal | 30m |
Claim: ROI for reliability is not found in headline metrics but in concrete changes to people and process. Quantitative measures like MTTR is misleading and… | ||
| Building and Running a Diversity-focused Pre-internship Program for SRE | Andrew Ryan | 28m |
12-week remote pre-internship trains about 96 students with Linux Foundation content and Major League Hacking administration, combines weekly mock interviews… | ||
| A Postmortem of SRE Interviewing | Michael Kehoe | 33m |
Applied to 37 companies and runs 113 interviews over 2.5 months, getting 12 non‑responses, 12 on‑sites and 8 offers; recommends clear timelines, standardized… | ||
| Self-Destructing Feature Flags | Jamie Gaskins | 29m |
Deploys include a runtime guard that watches errors and flips flags off automatically: a Ruby/Go prototype uses Redis-backed counters to track checks and… | ||
| Tales from the VOID: The Scary Truth about Incident Metrics | Courtney Nash | 27m |
An open database called the VOID collects nearly 2,000 reports for 660 organizations (status pages, blogs, postmortems) and finds duration data are strongly… | ||
| How We Survived (and Thrived) During The Pandemic and Helped Millions... | Chinmay Tripathi | 32m |
Company runs 80+ dev teams, 400+ apps and 100s of services across 150+ AWS accounts and 4,000 EC2 instances and absorbs a pandemic-driven spike with 100k new… | ||
| The Pandemic and The Classroom—Enabling Education for Millions | Alper Selcuk | 16m |
Describes concrete changes and results: tenant onboarding throughput increases 3x using parallel calls and multiple sync profiles, brownouts shave 30% off peak… | ||
| Applied Science Fiction: Operating a Research-Led Product | Noah Kantrowitz | 17m |
Research-led AR pipelines use Buildkite + Kubernetes job CI to run unit and slow functional tests, with CI doing 5–10 canned work units and a weekly 14-hour… | ||
| Taking the 737 to the Max | Nickolas Means | 34m |
Clear timeline of engineering and business choices that create failure: Boeing chooses to keep a common type rating and squeeze larger CFM LEAP engines onto an… | ||
| Securing Your Software Delivery Chain with Process Auditing | Shaun Mouton | 23m |
Presentation demonstrates using strace to record syscalls, file accesses and forks to build a baseline profile of a legacy app and using eBPF for low-overhead… | ||
| The Future of above-the-line Tooling | Richard I. Cook, John Allspaw | 28m |
Claim: tooling that represents and coordinates human work above the line is currently sparse but expanding rapidly, while classic observability tools like… | ||
| Tracing Bare Metal with OpenTelemetry | Amy Tobey, Shelby Spees | 32m |
Equinix Metal instruments the bare‑metal provisioning workflow end‑to‑end with OpenTelemetry, auto‑instrumentation and a custom tracing initializer plus an… | ||
| Are We There Yet? Metrics-Driven Prioritization for Your Reliability Roadmap | Christina Tan, Mindy Stevenson | 29m |
A metric-driven framework that maps business needs to focused SRE work and a compact reliability dashboard that rolls up SLIs/SLOs, MTTD, and change metrics… | ||
| SRE stands for...Skydiving Resilience Engineer | Victor Lei | 21m |
Skydiving statistics show about 1 in 500,000 fatality across ~3 million jumps per year, and safety relies on redundancy: main + reserve parachute, automatic… | ||
| Building a Path to the Future: Mentoring New SREs | Chastity Blackwell | 28m |
Focuses on mentoring early-career SREs with concrete practices: documents the why as well as the how, keeps a regular cadence of office hours, pairs by letting… | ||
| eBPF: The Next Power Tool of SREs | Michael Kehoe | 32m |
eBPF is presented as a kernel sandbox that SREs can use for high‑fidelity observability, networking and security by compiling C with clang/LLVM, loading via… | ||
| How the Metrics Backend Works at Datadog | Adam Mckaig, Tahia Khan | 31m |
Datadog’s metrics backend ingests tens of GB/s and billions of timeseries into PiBs of hot and cold storage, using agents, Redis intake queues, Cassandra for… | ||
| Automated Operating System and Environment Certification at LinkedIn... | Adam Debus | 31m |
LinkedIn builds an automation platform that cuts OS snapshot certification time by 70% faster in nine months, adds a statistical analysis engine to replace… | ||
| Triaging Real-time Security Threats with eBPF-powered Observability | Daniel Kim, Robert Prast | 26m |
Claims that Pixie + eBPF gives kernel-level, agentless observability that shortens triage cycles and removes guesswork; it captures raw DNS queries, full HTTP… | ||
| Exemplars in Practice: Finding the Needle in Your Observability Haystack | Gibbs Cullen | 24m |
Exemplars link a metric point to a single trace so engineers can jump from dashboards to traces but they fail when requests are heterogeneous or multi… | ||
| Dark Sky Camping: Reducing Alert Pollution with Modern Observability Practices | Kristin Smith | 29m |
Company scales from 100→1,000 parks (10x) in 18 months and faces alert chaos. Team moves from Prometheus+fragmented alerts into Honeycomb with the… | ||
| Ten-year Journey to 10,000 Production Machines | Rob Hirschfeld | 34m |
Claim of running well over 10,000 production machines, proven by a decade of re‑architectures: starts as a Chef-based MVP with a pixieboot loop, moves to a… | ||
| Beyond Distributed Tracing | Kyusoon Lee | 14m |
Introduces CUI attribution as a propagated metadata string (e.g. youtube/watch) built on Google Census’ baggage mechanism to aggregate latency and error… | ||
| History-based Latency Prober Tuning | Jeff Borwey | 12m |
Uses historical prober data to auto-tune alert thresholds by fitting log-normal fits or a skew-adjusted boxplot (metacouple skewness) and backtesting against… | ||
| Using Serverless Functions for Real-time Observability | Liz Fong Jones, Jessica Kerr | 32m |
A sub-second query engine at honeycomb.io moves from RAM to local SSD to months of S3-backed segments, then parallelizes heavy queries by spawning 30,000… | ||
| Improving How We Observe Our Observability Data: Techniques for SREs | Dan Shoop | 26m |
Argues that clearer observability comes from combining related metrics into single visuals rather than dumping dozens of separate charts, demonstrating a… | ||
| Principled Performance Analytics | Narayan Desai, Brent Bryan | 01h 10m |
An analytics method called Two Sigma partitions workloads into cohort-based baselines, fits per-cohort distributions (log-normal or normal approximations), and… | ||
| Modeling Alert Quality | Moshe Zadka | 29m |
Defines alert quality as three concrete classes—true alarms, false alarms, and missing alarms—and models it with measurable intervals: start-to-detect… | ||
| Emergent Organizational Failure: Five Disconnections | Mattie Toia | 29m |
Lists five disconnections that derail reliability: assuming a single “production” boundary (example: a deployment workstation reboot causing a 20‑minute… | ||
| DO, RE, Me: Measuring the Effectiveness of Site Reliability Engineering | Dave Stanke | 21m |
DORA adds SRE-specific survey items (a “duck‑typing” checklist) and finds a small majority of teams practice SRE to some degree; SRE adoption predicts higher… | ||
| The Scientific Method for Resilience | Christina Yakomin | 13m |
Uses a one-hour Failure Modes and Effects Analysis meeting to turn failure questions into explicit hypotheses (example: if the database dies, writes fail and… | ||
| A Fresh Look at Operational Debt | David Owczarek | 16m |
Defines operational debt = work to fix process gaps that present risk and links it to measurable impact on time‑to‑restore, using runbooks and API contracts as… | ||
| nothing matches | ||