microsoft
diff --git a/‎.gitattributes‎
Lines changed: 2 additions & 0 deletions b/‎.gitattributes‎
Lines changed: 2 additions & 0 deletions
diff --git a/‎.github/policies/milestoneAssignment.prClosed.yml‎
Lines changed: 4 additions & 4 deletions b/‎.github/policies/milestoneAssignment.prClosed.yml‎
Lines changed: 4 additions & 4 deletions
diff --git a/‎.github/skills/startup-perf/SKILL.md‎
Lines changed: 193 additions & 0 deletions b/‎.github/skills/startup-perf/SKILL.md‎
Lines changed: 193 additions & 0 deletions
diff --git a/‎.github/workflows/README.md‎
Lines changed: 23 additions & 0 deletions b/‎.github/workflows/README.md‎
Lines changed: 23 additions & 0 deletions
@@ -60,3 +60,5 @@
 # https://github.com/github/linguist/issues/1626#issuecomment-401442069
 # this only affects the repo's language statistics
 *.h linguist-language=C
+
+.github/workflows/*.lock.yml linguist-generated=true merge=ours
@@ -16,16 +16,16 @@ configuration:
           branch: main
       then:
       - addMilestone:
-          milestone: 13.2
+          milestone: 13.3
       description: '[Milestone Assignments] Assign Milestone to PRs merged to the `main` branch'
     - if:
       - payloadType: Pull_Request
       - isAction:
           action: Closed
       - targetsBranch:
-          branch: release/13.1
+          branch: release/13.2
       then:
       - removeMilestone
       - addMilestone:
-          milestone: 13.1.1
-      description: '[Milestone Assignments] Assign Milestone to PRs merged to release/13.1 branch'
+          milestone: 13.2
+      description: '[Milestone Assignments] Assign Milestone to PRs merged to release/13.2 branch'
@@ -0,0 +1,193 @@
+---
+name: startup-perf
+description: Measures Aspire application startup performance using dotnet-trace and the TraceAnalyzer tool. Use this when asked to measure impact of a code change on Aspire application startup performance.
+---
+
+# Aspire Startup Performance Measurement
+
+This skill provides patterns and practices for measuring .NET Aspire application startup performance using the `Measure-StartupPerformance.ps1` script and the companion `TraceAnalyzer` tool.
+
+## Overview
+
+The startup performance tooling collects `dotnet-trace` traces from an Aspire AppHost application and computes the startup duration from `AspireEventSource` events. Specifically, it measures the time between the `DcpModelCreationStart` (event ID 17) and `DcpModelCreationStop` (event ID 18) events emitted by the `Microsoft-Aspire-Hosting` EventSource provider.
+
+**Script Location**: `tools/perf/Measure-StartupPerformance.ps1`
+**TraceAnalyzer Location**: `tools/perf/TraceAnalyzer/`
+**Documentation**: `docs/getting-perf-traces.md`
+
+## Prerequisites
+
+- PowerShell 7+
+- `dotnet-trace` global tool (`dotnet tool install -g dotnet-trace`)
+- .NET SDK (restored via `./restore.cmd` or `./restore.sh`)
+
+## Quick Start
+
+### Single Measurement
+
+```powershell
+# From repository root — measures the default TestShop.AppHost
+.\tools\perf\Measure-StartupPerformance.ps1
+```
+
+### Multiple Iterations with Statistics
+
+```powershell
+.\tools\perf\Measure-StartupPerformance.ps1 -Iterations 5
+```
+
+### Custom Project
+
+```powershell
+.\tools\perf\Measure-StartupPerformance.ps1 -ProjectPath "path\to\MyApp.AppHost.csproj" -Iterations 3
+```
+
+### Preserve Traces for Manual Analysis
+
+```powershell
+.\tools\perf\Measure-StartupPerformance.ps1 -Iterations 3 -PreserveTraces -TraceOutputDirectory "C:\traces"
+```
+
+### Verbose Output
+
+```powershell
+.\tools\perf\Measure-StartupPerformance.ps1 -Verbose
+```
+
+## Parameters
+
+| Parameter | Default | Description |
+|-----------|---------|-------------|
+| `ProjectPath` | TestShop.AppHost | Path to the AppHost `.csproj` to measure |
+| `Iterations` | 1 | Number of measurement runs (1–100) |
+| `PreserveTraces` | `$false` | Keep `.nettrace` files after analysis |
+| `TraceOutputDirectory` | temp folder | Directory for preserved trace files |
+| `SkipBuild` | `$false` | Skip `dotnet build` before running |
+| `TraceDurationSeconds` | 60 | Maximum trace collection time (1–86400) |
+| `PauseBetweenIterationsSeconds` | 45 | Pause between iterations (0–3600) |
+| `Verbose` | `$false` | Show detailed output |
+
+## How It Works
+
+The script follows this sequence:
+
+1. **Prerequisites check** — Verifies `dotnet-trace` is installed and the project exists.
+2. **Build** — Builds the AppHost project in Release configuration (unless `-SkipBuild`).
+3. **Build TraceAnalyzer** — Builds the companion `tools/perf/TraceAnalyzer` project.
+4. **For each iteration:**
+   a. Locates the compiled executable (Arcade-style or traditional output paths).
+   b. Reads `launchSettings.json` for environment variables.
+   c. Launches the AppHost as a separate process.
+   d. Attaches `dotnet-trace` to the running process with the `Microsoft-Aspire-Hosting` provider.
+   e. Waits for the trace to complete (duration timeout or process exit).
+   f. Runs the TraceAnalyzer to extract the startup duration from the `.nettrace` file.
+   g. Cleans up processes.
+5. **Reports results** — Prints per-iteration times and statistics (min, max, average, std dev).
+
+## TraceAnalyzer Tool
+
+The `tools/perf/TraceAnalyzer` is a small .NET console app that parses `.nettrace` files using the `Microsoft.Diagnostics.Tracing.TraceEvent` library.
+
+### What It Does
+
+- Opens the `.nettrace` file with `EventPipeEventSource`
+- Listens for events from the `Microsoft-Aspire-Hosting` provider
+- Extracts timestamps for `DcpModelCreationStart` (ID 17) and `DcpModelCreationStop` (ID 18)
+- Outputs the duration in milliseconds (or `"null"` if events are not found)
+
+### Standalone Usage
+
+```bash
+dotnet run --project tools/perf/TraceAnalyzer -c Release -- <path-to-nettrace-file>
+```
+
+## Understanding Output
+
+### Successful Run
+
+```
+==================================================
+ Aspire Startup Performance Measurement
+==================================================
+
+Project: TestShop.AppHost
+Iterations: 3
+...
+
+Iteration 1
+----------------------------------------
+Starting TestShop.AppHost...
+Attaching trace collection to PID 12345...
+Collecting performance trace...
+Trace collection completed.
+Analyzing trace: ...
+Startup time: 1234.56 ms
+
+...
+
+==================================================
+ Results Summary
+==================================================
+
+Iteration StartupTimeMs
+--------- -------------
+        1       1234.56
+        2       1189.23
+        3       1201.45
+
+Statistics:
+  Successful iterations: 3 / 3
+  Minimum: 1189.23 ms
+  Maximum: 1234.56 ms
+  Average: 1208.41 ms
+  Std Dev: 18.92 ms
+```
+
+### Common Issues
+
+| Symptom | Cause | Fix |
+|---------|-------|-----|
+| `dotnet-trace is not installed` | Missing global tool | Run `dotnet tool install -g dotnet-trace` |
+| `Could not find compiled executable` | Project not built | Remove `-SkipBuild` or build manually |
+| `Could not find DcpModelCreation events` | Trace too short or events not emitted | Increase `-TraceDurationSeconds` |
+| `Application exited immediately` | App crash on startup | Check app logs, ensure dependencies are available |
+| `dotnet-trace exited with code != 0` | Trace collection error | Check verbose output; trace file may still be valid |
+
+## Comparing Before/After Performance
+
+To measure the impact of a code change:
+
+```powershell
+# 1. Measure baseline (on main branch)
+git checkout main
+.\tools\perf\Measure-StartupPerformance.ps1 -Iterations 5 -PreserveTraces -TraceOutputDirectory "C:\traces\baseline"
+
+# 2. Measure with changes
+git checkout my-feature-branch
+.\tools\perf\Measure-StartupPerformance.ps1 -Iterations 5 -PreserveTraces -TraceOutputDirectory "C:\traces\feature"
+
+# 3. Compare the reported averages and std devs
+```
+
+Use enough iterations (5+) and a consistent pause between iterations for reliable comparisons.
+
+## Collecting Traces for Manual Analysis
+
+If you need to inspect trace files manually (e.g., in PerfView or Visual Studio):
+
+```powershell
+.\tools\perf\Measure-StartupPerformance.ps1 -PreserveTraces -TraceOutputDirectory "C:\my-traces"
+```
+
+See `docs/getting-perf-traces.md` for guidance on analyzing traces with PerfView or `dotnet trace report`.
+
+## EventSource Provider Details
+
+The `Microsoft-Aspire-Hosting` EventSource emits events for key Aspire lifecycle milestones. The startup performance script focuses on:
+
+| Event ID | Event Name | Description |
+|----------|------------|-------------|
+| 17 | `DcpModelCreationStart` | Marks the beginning of DCP model creation |
+| 18 | `DcpModelCreationStop` | Marks the completion of DCP model creation |
+
+The measured startup time is the wall-clock difference between these two events, representing the time to create all application services and supporting dependencies.
@@ -99,3 +99,26 @@ When you comment on a PR (not an issue), the workflow will automatically push ch
 ### Concurrency
 
 The workflow uses concurrency groups based on the issue/PR number to prevent race conditions when multiple commands are issued on the same issue.
+
+## Backmerge Release Workflow
+
+The `backmerge-release.yml` workflow automatically creates PRs to merge changes from `release/13.2` back into `main`.
+
+### Schedule
+
+Runs daily at 00:00 UTC (4pm PT during standard time, 5pm PT during daylight saving time). Can also be triggered manually via `workflow_dispatch`.
+
+### Behavior
+
+1. **Change Detection**: Checks if `release/13.2` has commits not in `main`
+2. **PR Creation**: If changes exist, creates a PR to merge `release/13.2` → `main`
+3. **Auto-merge**: Enables GitHub's auto-merge feature, so the PR merges automatically once approved
+4. **Conflict Handling**: If merge conflicts occur, creates an issue instead of a PR
+
+### Assignees
+
+PRs and conflict issues are automatically assigned to @joperezr and @radical.
+
+### Manual Trigger
+
+To trigger manually, go to Actions → "Backmerge Release to Main" → "Run workflow".