Recommended reading
Bookmark these. Read them before the incident, not during it — a team fluent in this material diagnoses .NET performance issues far faster, and most of these tutorials map almost one-to-one onto the problems above.
Start with the Virto Commerce material:
- Virto Commerce Production Checklist - howto - Virto Commerce
- Do Not Blindly Trust Application Insights Durations When CPU Is Overloaded — why durations lie under CPU pressure. The single most important read before you interpret any AI chart.
- How to Contribute to Virto Commerce — when you find the fix, ship it.
- Capture web session traffic
- Virto Commerce Docs Are Now Available in Context7 — point your AI assistant at the current docs.
Then the Microsoft .NET documentation. This is the canonical, vendor-grade reference for everything we discussed — and it’s free:
- Usage analysis with Application Insights - Azure Monitor | Microsoft Learn
Application Insights OpenTelemetry observability overview - Azure Monitor | Microsoft Learn - .NET diagnostics tools overview — the map. Debuggers, profilers, and the three pillars of observability (metrics, logs, traces). Start here to understand which tool fits which symptom.
- Tutorial: Debug high CPU usage in .NET — a hands-on walkthrough using
dotnet-countersto confirm high CPU anddotnet-traceto capture a flamegraph of the hot path. This is exactly the non-prod investigation we ask for after scaling. - Performance diagnostics tools for ASP.NET Core — Visual Studio profilers, Application Insights, and PerfView in an ASP.NET Core context (which is what Virto Commerce is).
- Guidance for troubleshooting high CPU usage — the broader methodology for sustained ≥80% CPU, including data collection with ProcDump and WPR.
Please feel free to share your references and thoughts.