Profiling data with Pat Somaru and Narayan Desai

In the opening episode of the third season of Google's Site Reliability Engineering (SRE) podcast, the focus shifts to the changing landscape of software engineering within SRE practices. Hosts engage with industry experts Narayan Desai and Pat Somaru to discuss how to enhance observability beyond the usual metrics, logs, and traces. They tackle the challenges of turning complex data into actionable insights, touching on topics like mental models, workload context, and the nuances of setting effective Service Level Objectives (SLOs). By examining the limitations of current systems and promoting simpler, more effective approaches, the episode highlights the need for organizations of all sizes to make observability both accessible and relevant.

Outlines

Sign in to continue reading, translating and more.

Continue

Google SRE Prodcast

Season 3 Introduction and Guest Introductions

Beyond the Three Pillars of Observability

The Importance of Simple Insights from Complex Data

Bridging Workstation Profiling to Distributed Systems Observability

Mental Models, Noise Reduction, and Workload Modeling

Workload Context and the Limitations of Averaging Metrics

Addressing Challenges in SLOs and the Importance of Variance Modeling

Needle-in-a-Haystack Problems and High-Cardinality Insights

Observability for Smaller Organizations and Practical Considerations

Common Fallacies in Observability and the Importance of Context

Profiling data with Pat Somaru and Narayan Desai

Google SRE Prodcast

00:01Season 3 Introduction and Guest Introductions

Season 3 Introduction and Guest Introductions

01:09Beyond the Three Pillars of Observability

Beyond the Three Pillars of Observability

03:24The Importance of Simple Insights from Complex Data

The Importance of Simple Insights from Complex Data

06:51Bridging Workstation Profiling to Distributed Systems Observability

Bridging Workstation Profiling to Distributed Systems Observability

09:13Mental Models, Noise Reduction, and Workload Modeling

Mental Models, Noise Reduction, and Workload Modeling

14:47Workload Context and the Limitations of Averaging Metrics

Workload Context and the Limitations of Averaging Metrics

19:38Addressing Challenges in SLOs and the Importance of Variance Modeling

Addressing Challenges in SLOs and the Importance of Variance Modeling

24:44Needle-in-a-Haystack Problems and High-Cardinality Insights

Needle-in-a-Haystack Problems and High-Cardinality Insights

26:55Observability for Smaller Organizations and Practical Considerations

Observability for Smaller Organizations and Practical Considerations

30:09Common Fallacies in Observability and the Importance of Context

Common Fallacies in Observability and the Importance of Context