Operations | Monitoring | ITSM | DevOps | Cloud

In the September edition of the Kubernetes Monitoring Helm chart office hours, we discuss the version 4.4 and 4.5 releases, the upcoming 4.6 feature release, and the next 5.0 major release.

Azure architecture diagrams go out of date fast, and management-plane data alone misses the runtime connections that matter most. In v5.4, diagrams move to their own Diagrams tab in Azure Documenter. Alongside the enhanced network and workload diagrams, there is a new Resource Visualizer diagram. Scope it by subscription and resource group, or write your own custom Azure Resource Graph query to define exactly which resources to include.

"Our Azure bill went up 20% last month." Without context, finance only sees a rising cost. Unit economics gives them the full picture. Turbo360 lets you overlay business KPIs on your Azure spend. Track units like orders, active users, document views, or monthly recurring revenue alongside cost, and see your cost per unit month over month. Now the conversation becomes: "Orders went up 150% and our cost per order came down." That is a story about efficiency, not overspend.

You wouldn’t ship untested SQL Server database changes – Why is Databricks different? Databricks is where the data estate is growing, and increasingly where AI workloads run and generate change. But schema change there still happens the hard way: views and stored procedures managed through manually versioned scripts, drift between workspaces discovered when a deployment fails, and no reliable record of what changed, when, where, or why. As AI raises the volume and speed of schema change, these gaps widen.

Most teams collect the profiles and traffic data that explain a slowdown. Almost nobody has time to read it before users notice. In this Product Highlights conversation, Sylvain Guittard, Senior Director of Product at Upsun who leads the team behind the Upsun console and CLI, breaks down the Upsun Cloud Performance Agent, the first background agent running on Upsun Cloud. His take: "We monitor everything, we feed that into an agent, and the agent will be capable of finding what the bottlenecks are in your application. And on top of it, it gives you a patch, or a way to fix it." We get into.