关于IIS 8.5中COM+应用性能监控指标及PAL工具的问询

Great question—let’s break this down step by step, covering the metric applicability, PAL tool updates, and some hands-on tips from recent work monitoring legacy COM+ and ASP apps on IIS 8.5.

Are the Listed Metrics Still Valid for IIS 8.5? #

Absolutely—most of these metrics remain fully relevant and actionable for monitoring ASP-based COM+ applications on IIS 8.5. Here’s why each set matters:

• ASP Error Metrics: Errors During Script Runtime, Errors From ASP Preprocessor, and Errors From Script Compilers are all tracked by IIS 8.5’s classic ASP performance counters (just make sure the ASP feature is enabled in Server Manager). These are critical for catching script-level failures that might tie to COM+ integration issues.
• Request & Session Metrics: Requests Executing, Requests Queued, Sessions Total, Requests/Sec, Requests Total, Requests succeeded, and Requests Failed Total are core to IIS request handling. Classic ASP workloads rely on these counters to spot traffic spikes, queue bottlenecks, and overall request health—no meaningful changes here for IIS 8.5.
• Timing Metrics: Request Execution Time and Request Wait Time are still key for identifying latency. For COM+ apps, these can help you distinguish between slow ASP script execution vs. delays in COM+ method calls (a common pain point).
• Cache & Debugging: Template Cache Hit Rate tracks how often IIS reuses compiled ASP templates (critical for optimizing script performance), and Debugging Requests monitors active ASP debug sessions—both are still supported in IIS 8.5.
• Process Memory: Process (inetinfo) Private Bytes is valid, but a quick note: in IIS 7+, most application pools run in w3wp.exe (even for classic ASP). If your COM+ app uses a classic mode pool, inetinfo might still host some components, but I’d recommend monitoring w3wp.exe memory as well to cover all bases.

PAL Tool Migration to Git #

You’re correct that the PAL (Performance Analysis of Logs) tool moved from CodePlex to Git. Good news: it’s still actively maintained, fully compatible with IIS 8.5 PerfMon logs, and can parse all the metrics you listed. It generates detailed reports that highlight anomalies like high queue lengths, slow request times, or memory leaks—super useful for turning raw PerfMon data into actionable fixes.

Hands-On Tips from Recent COM+ Monitoring Experience #

From my work supporting legacy COM+ and ASP apps on IIS 8.5, here are a few extra pointers:

• Add COM+-Specific Counters: Don’t stop at IIS/ASP metrics. Include COM+ counters like Calls/sec, Queue Length, Method Call Duration, and Activation Failures—these directly tell you if the COM+ component itself is the bottleneck (not just the ASP layer).
• Baseline First: Capture a baseline of metrics during normal operation, then compare it to logs during slowdowns. PAL’s comparison reports make it easy to spot sudden spikes (e.g., a jump in Requests Queued that coincides with COM+ activation failures).
• Combine with Failed Request Tracing: For specific failed requests, pair PerfMon data with IIS Failed Request Tracing. This lets you drill into exactly where the error occurs—whether it’s in the ASP script, a COM+ method call, or IIS configuration.
• Watch for COM+ Memory Leaks: COM+ components can leak memory over time. Track Process (dllhost.exe) Private Bytes if your COM+ app runs in a separate process—this is often overlooked but critical for long-term stability.

内容的提问来源于stack exchange，提问作者Dustin Eastman