SRE NY Tech Talks

Google SRE proudly announces the next event in its Site Reliability Engineering (SRE) Tech Talks series on Wednesday, August 2nd at Google’s Pier 57 building in NYC. The event starts at 6:00 PM and lasts until 8:30 PM.

We invite you to join us for an hour of short talks on Reliability and DevOps topics, followed by an opportunity to meet and talk with fellow engineers over light refreshments. We are pleased to welcome the following speakers:

• Jeff Luery and Yash Mestry, Perpetual. “DevSecOps and SRE integration”
  Our talk will specifically communicate how SRE and DevOps processes can be implemented into enterprise software engineering and web development projects. We will cover industry best practices, as well as magical tools and tricks to maximize server uptime, performance, reliability and overall efficiency.
• Hasit Mistry, FluxNinja: “Achieving Fault Tolerance with Observability-driven Load Management”
  This talk will help the audience build an intuition about load management, starting from basic principles of queuing theory and Little's law. These principles help build understanding of complex failure scenarios and how they manifest in microservices. Following this, the session illustrates how early adopters are making use of Aperture to gracefully degrade their applications during complex failures. In essence, the goal of this talk is to enhance the community's collective understanding of system reliability and the potential of observability driven closed loop automation techniques for effective load management. Ensuring reliable operation of microservices is a challenging task. Metastable failures such as cascading overloads, retry storms and death spirals cause services to enter a permanent state of failure that requires manual intervention to recover. Mitigation strategies like circuit breakers and auto-scaling fall short due to their narrow vantage points.
  To operate microservices reliably, observability-driven automation is required. Aperture is an open source load management system that leverages CNCF technologies such as etcd, Prometheus, OpenTelemetry, Open Policy Agent, and Istio/Envoy. It combines ideas from the world of observability, control systems, and network scheduling to automate service protection and workload prioritization.
• Andreas Bobak, Google NYC: “Frontend Design by SREs for SREs”

In a world where SRE is quickly changing from running and maintaining their own “scripts” and writing large applications so that others can maintain and monitor their systems, new challenges wrapping up their tooling into something that low-context users can easily utilize. With that a whole new world of SRE frontend emerges. Here are a few ways SREs can think about doing good user interface design for their user-journeys.