SREDAY

Inspired by work from hyperscale SRE teams, RunWhen is building the industry's largest open source library of troubleshooting automation. Contributing Authors receive royalties when the company's enterprise customers use their code to automate root cause analysis and remediation.

Unikernels have been running in production for years now, and that experience has produced a set of new tools, best practices, and workflows.In this talk, come hear about how and why your peers have been pushing unikernels to production and how new methods have been established. Whether it's cloud-specific monitoring, accessing GPU resources for ML workloads, or applying security policies that Linux workloads can't, we'll show you how companies, large and small, have been heavily investing in unikernel deployment.

Test-Driven Development (TDD) gives you the peace of mind to sleep well, knowing your code is robust and reliable. By writing tests before the implementation, you ensure that each change meets the requirements and works as intended, reducing the anxiety of unexpected bugs. TDD also enhances the requirements stage, as it forces clear, precise specifications, leading to a more predictable and controlled development process.

Sebastian is the CTO and Co-founder of Codejet. With 8 years of experience as a software architect and lecturer, he has worked at companies like Goldman Sachs, where he played a key role in developing Marcus.com. He also contributed to Stageclip.com and AndroidPolice.com. Outside of work, Sebastian enjoys Brazilian Jiu-Jitsu and motorcycles.

Is the app slow or the network lagging? When it comes to latency across distributed systems, it can be hard to pinpoint where exactly the issue lies. To add to this, business demands drive where we run our workloads today - whether on-premises, cloud or hybrid environments - to enhance agility, reduce cost or risk, but often blurring our view into the health of our systems across these disjoint environments. Clock synchronization in distributed systems is key for identifying the true culprit when it comes to latency. In this session, I’ll demystify the building blocks for synchronizing clocks across distributed systems. I’ll go over protocols to date (NTP, PTP, PPS), their pros and cons, and best practices such as slewing clocks. I’ll dive into an algorithm that achieves higher precision clock synchronization at scale. Using this, we gain visibility into useful metrics such as one way delays, which gives us the ability to monitor response times and improve performance for our end users. The session is targeted for SRE’s, infrastructure, and platform teams who are looking to gain a deeper understanding of how to improve response times for workloads with high precision clock synchronization across distributed systems deployed on-premises, in the cloud, multi-cloud, or hybrid environments.

Lerna is a Senior Solutions Engineer at Clockwork Systems where she helps customers meet their performance goals as well as regulatory reporting objectives with software solutions built upon Clockwork.io’s foundational clock synchronization research. Prior to this, she was an AWS Senior Solutions Architect serving Global Financial Services customers for 3 years. Before that, Lerna spent 17 years as an infrastructure engineer in large financial services companies working on authentication systems, distributed caching, and multi region deployments using IaC and CI/CD to name a few. In her spare time, she enjoys hiking, sight seeing and backyard astronomy.

With Infrastructure as Code (IaC) becoming the de facto way we manage our infrastructure today, a lot of excellent tools have become widely adopted that each have a different set of strengths. In this talk, we'd like to take a look at the evolution of the IaC landscape over the past decade and where we're heading, with a focus on Site Reliability Engineering (SRE) and modern cloud-native operations challenges. We'll examine some of the biggest SRE and operations challenges through the lens of IaC, including disaster recovery, security, cost optimization, performance tuning, and even where complexity factors in when choosing your tooling of choice. While many teams have already selected their IaC tooling, we may also like to consider migration and integration of multiple tools for different use cases and stacks. In this interactive talk, we'll allow you to decide which tools we explore - from CDK to Pulumi, Terraform, OpenTofu, Helm, ArgoCD - to learn about how they stack up versus modern cloud-native SRE challenges.

Ohad Maislish is the CEO and co-founder of env0 and part of the founding team for the OpenTofu project. Before env0, Ohad was the CEO and founder of Arno Software, a cloud infrastructure services company, and Capester, a startup that empowered citizens in smart cities worldwide. Over the course of this career, Ohad has also served in different technical and management roles at Ravello Systems, eToro, and VMware. He was also the youngest developer at Microsoft Israel at the age of 17, after starting his bachelor's degree at the age of 14.

When migrating to Kubernetes from legacy architecture, organizations typically see application availability suffer. That wasn’t the case though for Acquia, whose website hosting platform supports tens of thousands of highly variable workloads. In this session, we’ll see how Acquia overcame the reliability and cost challenges of migrating to Amazon EKS and Kubernetes by leveraging machine learning and automation to rightsize their workloads, achieving 99.99% uptime.

Presented by: Erwin Daria, Director of Alliances at StormForge

Erwin Daria is the Director of Alliances at StormForge. Over his 25 year career, Erwin has served in roles spanning from enterprise infrastructure engineering to sales engineering to product management and everything in-between for companies like Medivation, Tintri, and Juniper Networks. Erwin’s professional passions lie in building companies and technology that profoundly improve operational efficiency and productivity.

Modern software is all but delivering moments of delight to customers. Moments that are constantly competing for users attention , reliability is not just a luxury—it’s a necessity. But what does reliability truly mean for stakeholders, and how do they ensure it in our software systems? This talk explores the evolution of reliability from the perspective of someone who has navigated the trenches of shipping reliable products, managing large SRE teams, and building an organization dedicated to SRE platforms.

Reliability in software is a multifaceted journey that begins with understanding when and why it becomes critical. This talk will address key milestones in this journey:

1. When Does Software Require Reliability?
2. Understanding the tipping points where reliability moves from being an optional feature to a core necessity.
3. Identifying the stakeholders who benefit the most from reliable systems.
4. Investing in Observability: The Next Step in Reliability
5. Exploring the reasons why reliability demands investment in observability.
6. Discussing the core components of observability: logs, metrics, and traces.
7. Real-world examples of how observability transforms reliability.
8. From Observability to Controllability: Bridging the Gap
9. Defining controllability and its significance in maintaining reliable systems.
10. How observability enables controllability and proactive management.
11. Practitioner Needs vs. Jargon Overload
12. Cutting through the noise: Focusing on the practical needs of practitioners rather than getting lost in the technical jargon.
13. Sharing insights and lessons learned from real-world experiences in SRE and product management.
14. Building a Reliable Organization
15. Strategies for fostering a culture of reliability within teams.
16. Tools and practices that support reliable software delivery.

Everyone talks about how to do SRE. Here, we talk about "why to do SRE at all"

Time to wrap SREday SF - where we go next, and how you can get involved

SREs have been at the forefront of Incident management when it comes to recovering the services during the outages. While SREs can follow the best practices such as setting up the SLOs and auto recovery systems as much as possible, the P0 and P1 outages require the developers to join the forces and recover the services in the quickest time possible. Efficient collaboration between SREs and Developers is found to be crucial in achieving successful incident resolution outcomes. This talk focuses on the typical challenges faced at both the Developer and the SRE teams in keeping the communication and technical knowledge sharing to the best of their ability. This talk also discusses some modern solutions such as measuring and improving the Incident Response Readiness. SREs will walk away with new insights on efficient developer collaboration.

Infrastructure breaks, but systems can persist! We really want systems that withstand unavoidable failures. Abstractly, we understand that. In this talk, Steve presents concepts like spanning failure domains and using generic mitigations, as well as introducing a lab environment where teams can experiment with these directly and a community where we can discuss it all, memes and all.

In Observability, Cardinality often gets the short end of the stick due to high costs. This forces engineers to minimize its use. However, SREs, who need to identify business impact quickly, suffer the most from this limitation because of misaligned incentives between developers & SREs. Cardinality reveals crucial business dimensions, linking technical metrics to business outcomes. Imagine if cost wasn't a concern and cardinality was basically free, you would now start seeing advanced tools emerge that could do the slicing on behalf of engineers. In this keynote, we will explore what this evolution in troubleshooting UX could look like?

Ruchir is the co-founder and CEO of Cardinal, which helps engineers view machine data through a business lens. In his previous life, he spent 7 years as a Lead Engineer on Netflix's Observability team, where he built petabyte-scale Observability products that are used daily by thousands of Netflix engineers.

Coming soon...