PSU App Randomly becomes unresponsive
It seems that upgrading to 5.6.0 fixed my issues. Specifically, I believe this bug fix. Environment Recycling · Issue #4703 · ironmansoftware/powershell-universal · GitHub
PSU App Randomly becomes unresponsive
Body:
I’m troubleshooting an intermittent issue with one of our production PowerShell Universal apps and would appreciate any insights or suggestions.
Environment
PowerShell Universal v5.5.2
Custom PowerShell 7 environment (pwsh 7.5)
Max runspaces set to 200 (configured in environment settings)
Unlimited virtual memory
Hosted on Windows Server 2022, using MSSQL and the built-in web server
4 CPU, 24GB RAMGenerally this stays about 5% cpu/35% memory usage
Issue
We have a PSU that exposes several Pages to manage mostly on-prem AD accounts, but there are also some Graph calls. It generally works fine, but 2–3 times per week, it becomes completely unresponsive. When this occurs:
The affected app won’t load or refresh
Other PSU apps and scripts remain fully functional
Restarting the app (via PSU admin interface) immediately resolves the issue
What I’ve Tried
Enabled the built-in “Excessive Runspace Usage” health check which never triggers (threshold set to 50)
No high CPU or RAM usage on the server during hangs
No logged errors or warnings around the time of freeze. Debug logging is enabled.
Audited and optimized AD queries:Replaced unconstrained Get-ADUser calls
Using -Filter * with -ResultSetSize and -ResultPageSize capped at 200
Verified via the Diagnostics → Runspaces view that the environment does spin up many idle runspaces (10+), so scaling appears functional
When the app is frozen, the Sessions list in the PSU Admin UI is empty, except for a possible Security runspace
Questions
Are there any known issues with runspace cleanup, orphaned sessions, or resource locking that might cause an app-level freeze?
Is it possible for a PSU App to exhaust or block all runspaces in a way that doesn’t trigger the built-in health check?
Could this be a case of thread pool starvation or something PSU 5.5.2 may have fixed/addressed in a later version?Bottom of Form