SREcon25 Americas

Todd Underwood, Anthropic, and Brendan Burns, Microsoft

Format: Breakout Group Discussion


Running ML systems is a major new area for many SRE organizations. This session will dive into the differences between running reliable software services in general and ML systems: infrastructure considerations, monitoring, rollouts, performance and cost management, and more.



https://www.usenix.org/conference/srecon25americas/presentation/discussion-ml-production

Niall Murphy, Stanza, and Kurt Andersen, Clari

Format: AMA Session


Are you confused by the alphabet soup: PRRs, ICS, MTTR, o11y, SLOs, SLIs? Is SRE the same thing as DevOps? Will doing everything the SRE books say lead to success and a good night's sleep oncall?


This discussion session is the place to bring any and every question you may have about getting started as an SRE. Expect breadth rather than depth.


https://www.usenix.org/conference/srecon25americas/presentation/discussion-intro-sre

David Woods, The Ohio State University

Format: AMA Session


Dr. David D. Woods is a cognitive psychologist and systems safety expert who has spent his career studying resilience in systems and how people and machines can best work together—topics deeply relevant to SRE. Dr Woods will host a wide-ranging discussion on how systems adapt in the face of disturbances, how complex systems break down, how the human perceptual system works with user interfaces, and more.


https://www.usenix.org/conference/srecon25americas/presentation/discussion-ama-woods

Beth Adele Long, Adaptive Capacity Labs, and Fred Hebert, Honeycomb.io

Format: Breakout Group Discussion


The focus of this session is developing individual contributor skills in SRE. SREs do a lot of different things, including but not limited to: load testing, setting up and maintaining infrastructure services, building integration test pipelines, setting SLOs, maintaining alerts, being an incident commander, writing post incident reviews, system design for scalability, building automation, contributing code changes to core products which are aimed at increasing reliability or performance, and troubleshooting. Few engineers come to SRE with all of these skills. How, as practitioners, should we think about building skills in new areas? Does it make sense to be an SRE jack-of-all-trades, or should one specialize?


https://www.usenix.org/conference/srecon25americas/presentation/discussion-sre-ics

Colette Alexander, HashiCorp, and Sarah Butt, Salesforce

Format: Breakout Group Discussion


This session sets out to explore SRE at the level of the engineering team. Most engineering teams do some form of planning and goal setting—is this process different for SRE teams compared to other engineering teams? Do OKRs even make sense for SREs? How does your SRE team onboard new members?


There are several team practices specific to SRE, including (but not limited to) Production Readiness Reviews, regular production meetings, and wheel of misfortune exercises. What works and what doesn't? Why? What team-level practices are most important?


https://www.usenix.org/conference/srecon25americas/presentation/discussion-sre-team-practices

Alex Hidalgo, Nobl9, and Cail Young, Octopus Deploy

Format: AMA Session


Service Level Objectives are a core element of many organizations' SRE strategies. SLOs seem simple, but the way in which they are constructed and used has broad consequences for organizations, and for SRE teams' success. The session hosts will lead a discussion that gets into the nitty-gritty of implementing, maintaining, and living with SLOs and error budgets.



https://www.usenix.org/conference/srecon25americas/presentation/discussion-slos

Yvonne Lam and Mike Rembetsy, Bloomberg

Format: Breakout Group Discussion


Technical debt—decisions made to deliver in the short term with costs in the longer term—has not traditionally been seen as an SRE concern, but it should be. SRE teams with responsibility for production services generally need to contend with technical debt in the software they create (monitoring, automation, and so on) as well as being on the sharp end of many of the consequences of technical debt in the services they run. What are your experiences of technical debt as an SRE? How have you managed to cope with or to reduce technical debt?


https://www.usenix.org/conference/srecon25americas/presentation/discussion-tech-debt

Daria Barteneva, Microsoft Azure

Format: Breakout Group Discussion


This discussion will cover all forms of observability (logs, metrics, distributed traces). How should we think about the goals of observability programs and how do we know if we are achieving good outcomes?


https://www.usenix.org/conference/srecon25americas/presentation/discussion-observability

Blake Bisset and Robert Barron, IBM

Format: Unconference Session


If you were disappointed to not see a particular topic on the discussion track agenda then this is the session for you.


Come to this session—which will be run using the Open Spaces Technology unconference format—and propose your topic. The four topic areas with the most interest will be selected for discussion in breakout groups.



https://www.usenix.org/conference/srecon25americas/presentation/discussion-open-unconference