aspire stop --all can fail after AppHost processes exit during termination monitoring

[davidfowl]

Summary

aspire stop --all can report successful stops for multiple AppHosts and then exit with an unexpected error while monitoring process termination.

Environment

• Aspire CLI version: 13.4.0+3f0998a95833d5fef96b3159d56b667889b71166
• Aspire CLI build ID: 13.400.26.27718
• OS: macOS
• Command: aspire stop --all

What happened

The command found six running AppHosts, sent stop signals, and printed several successful stop messages. It then failed with exit code 1.

Relevant log excerpt:

[2026-05-29 14:39:34.276] [DBUG] [StopCommand] Found 6 running AppHost(s) to stop
...
[2026-05-29 14:39:34.531] [INFO] [Stdout] ✅ FoundryAgents.AppHost.csproj stopped successfully.
[2026-05-29 14:39:34.531] [INFO] [Stdout] ✅ Aspire.Dev.AppHost.csproj stopped successfully.
[2026-05-29 14:39:34.531] [INFO] [Stdout] ✅ davidfowl-glowing-guide/pr-timeline-app.AppHost/pr-timeline-app.AppHost.csproj stopped successfully.
[2026-05-29 14:39:34.532] [INFO] [Stdout] ✅ davidfowl-cautious-goggles/pr-timeline-app.AppHost/pr-timeline-app.AppHost.csproj stopped successfully.
[2026-05-29 14:39:34.533] [INFO] [Stdout] ✅ davidfowl-bookish-umbrella/pr-timeline-app.AppHost/pr-timeline-app.AppHost.csproj stopped successfully.
[2026-05-29 14:39:34.534] [FAIL] [AspireCliTelemetry] An unexpected error occurred: Unable to retrieve the specified information about the process or thread.  It may have exited or may be privileged.
System.ComponentModel.Win32Exception (3): Unable to retrieve the specified information about the process or thread.  It may have exited or may be privileged.
   at System.Diagnostics.Process.get_StartTimeCore() + 0xf4
   at System.Diagnostics.Process.get_StartTime() + 0x20
   at ProcessSignaler.TryGetRunningProcess(Int32, Nullable`1, ILogger) + 0x5c
   at Aspire.Cli.Processes.ProcessShutdownService.IsProcessStopped(ProcessShutdownService.ProcessTarget) + 0x20
   at System.Linq.Enumerable.All[TSource](IEnumerable`1, Func`2) + 0x134
   at Aspire.Cli.Processes.ProcessShutdownService.<MonitorProcessesForTerminationAsync>d__17.MoveNext() + 0xe8
--- End of stack trace from previous location ---
   at Aspire.Cli.Processes.ProcessShutdownService.<StopProcessesAsync>d__11.MoveNext() + 0x104
--- End of stack trace from previous location ---
   at Aspire.Cli.Processes.ProcessShutdownService.<StopAppHostAsync>d__10.MoveNext() + 0xc8
--- End of stack trace from previous location ---
   at Aspire.Cli.Commands.StopCommand.<>c__DisplayClass17_0.<<StopAppHostAsync>b__0>d.MoveNext() + 0x70
--- End of stack trace from previous location ---
   at Aspire.Cli.Interaction.ConsoleInteractionService.<ShowStatusAsync>d__28`1.MoveNext() + 0x78
--- End of stack trace from previous location ---
   at Aspire.Cli.Commands.StopCommand.<StopAppHostAsync>d__17.MoveNext() + 0x130
--- End of stack trace from previous location ---
   at Aspire.Cli.Commands.StopCommand.<StopAllAppHostsAsync>d__16.MoveNext() + 0x160
--- End of stack trace from previous location ---
   at Aspire.Cli.Commands.StopCommand.<ExecuteAsync>d__13.MoveNext() + 0x220
--- End of stack trace from previous location ---
   at Aspire.Cli.Commands.BaseCommand.<>c__DisplayClass15_0.<<-ctor>b__0>d.MoveNext() + 0xc0
[2026-05-29 14:39:34.539] [INFO] [Stderr] ❌ An unexpected error occurred: Unable to retrieve the specified information about the process or thread.  It may have exited or may be privileged.
[2026-05-29 14:39:34.540] [INFO] [Cli] Exit code: 1

Expected behavior

If a process exits while aspire stop --all is monitoring termination, the CLI should treat that as stopped instead of surfacing a Win32Exception from Process.StartTime and failing the command.

[joperezr]

🤖 Aspire triage pass — 2026-05-29 15:18 UTC

Area: area-cli
Kind: bug (possible regression)
Repro: yes-large — stack trace and environment captured in the issue body (lines ~8–55)
Affected version(s): 13.4.0+3f0998a95833d5fef96b3159d56b667889b71166 (latest in-flight: 13.4)
Severity: S3 — AppHost processes do get stopped successfully, but the command surfaces a Win32Exception from Process.StartTime and exits 1. Workaround: ignore the exit code (the stop itself succeeded).

Urgency: 🟡 vNext — recommend milestone 13.5
Reason: S3 with effective workaround (processes are stopped). Cannot confirm with high confidence whether this is a strict regression in 13.4 vs. a pre-existing race in ProcessSignaler.TryGetRunningProcess, so per the 13.4 confidence rules this stays out of the snapped release.

Suggested owner(s): @danegsta (primary), backup @JamesNK
Evidence:

• @danegsta merged Fix aspire stop falsely reporting failure on Unix #17612 ("Fix aspire stop falsely reporting failure on Unix") on 2026-05-28, touching the exact files in this stack trace: src/Aspire.Cli/Processes/ProcessShutdownService.cs and src/Shared/ProcessSignaler.cs. This issue looks like a closely related code path (Process.StartTime throwing when the process has already exited mid-monitor).
• CODEOWNERS for /src/Aspire.Cli: @mitchdenny @JamesNK @davidfowl (author excluded).
• @JamesNK has overwhelming recent activity in area-cli (30+ merged PRs in the last 90 days), making them a strong backup.

Confidence: medium
Needs human?: no

Dry-run mode — no labels or milestone applied. Reply /triage apply to promote, or edit and reapply.

[davidfowl]

Fixed