About this event
Join fellow cloud architects and engineers for deep dives into topics related to Kubernetes and cloud-native computing!
- Title: OpenTelemetry and Continuous Feedback - Things you need to know about your code at runtime
- Presented by: Roni Dover, Co-founder and CTO at Digma
- Description: What do you know about the code changes that were just introduced into the codebase? When will you start noticing if something goes wrong? If there are so many accessible observability sources that can tell us what the code is doing, why are we using so little of that in our day-to-day coding?
Continuous Feedback is a new dev practice that aims to make practical usage of code runtime data to shorten the feedback loop during development. It enables developers to get early data about their code changes and detect issues and regressions as-they-code. At the same time, collecting data from multiple environments, allows developers to instantly understand how their code is performing in the real world.
In this session, we'll look past the novelty of using OSS observability tools and technologies, to discuss how we can actually make them useful for developers. We'll take a look at the benefits of enabling OpenTelemetry collection for dev and test data and examine OSS tools to help analyze the application runtime. Throughout the talk, we'll go over code examples of common anti-patterns, code smells, hidden errors, and other types of problems that observability can reveal - prior to merging a PR,
Ultimately, the goal should not be simply observing the application or creating nice-looking dashboards. Rather, success is in leveraging observability data in order to achieve a more effective dev process and write better code.
- Title: A bigger house requires a good Karpenter
- Presented by: Jeremy Cowan, Sr. Manager, Developer Advocacy at AWS
- Description: This talk will provide a brief overview of the Karpenter project, an open source cluster autoscaler for Kubernetes. It will also cover how the project is aligning with the CAS API in preparation for its donation to SIG autoscaling and newly released features, including Azure's provider for AKS and new controls for managing disruptions, e.g. during node upgrades and consolidation.
Karpenter is increasingly becoming the de facto way to scale Kubernetes clusters on AWS. Many are attracted to it because of its ability to simplify node management and its ability to binpack pods onto as few (least costly) instances as possible without affecting reliability or performance.