Troubleshooting¶

The app doesn't start¶

Port in use: another program uses the port. Change it with --urls.
Missing runtime: dotnet gView.Pipe.Shell.dll needs the .NET 10 ASP.NET Core runtime (not needed for Docker).
No write access: the account needs write access to the data folder.

The plugin isn't installed or couldn't be loaded. Check:

The folder is plugins/<Name>/ and contains <Name>.dll with exactly the folder's name.
The log at the start. A plugin that can't be loaded gives a warning that starts with Plugins:.
You restarted the app after copying the plugin.

The script isn't in a released folder, or the folder isn't in Pipe:Scripts:ScriptRoots.
The interpreter isn't installed (python, pwsh, bash). Set the path in Pipe:Scripts:Interpreters.
Run Scripts Repo: git isn't installed, the Git server or its credential is wrong, or the repository filter hides the repository. The node shows hints about missing settings.
In Docker: .ps1, .bat and .cmd scripts can't run in the Linux image.
The node reports failed although the script looks fine: a non-zero exit code or a line that matches the error pattern fails it. Look at the log of the node for the output of the script.

Answer	Cause
`404`	The workflow isn't active, or the path is wrong. Webhook paths are lower case.
`405`	The method differs from the one set in the trigger.
`401`	The credential is missing or wrong.
`413`	The body is larger than 10 MB.
`429`	The workflow is already running, its limit of concurrent runs is reached and no room is left in its queue.
`503`	The run waited in the queue longer than the webhook's timeout. Retry later, or raise the timeout or Max concurrent runs.
`500`	The run failed. The body has the error and the `executionId`; look the run up in the history. See Error handling.

If a workflow can't be activated because its path is already used, the editor shows which workflow and node use it.

The data/keys/ folder doesn't match data/credentials/. Restore both from the same backup, or enter the credentials again.

Runs that were in progress end with a restart. Active workflows are armed again, but a Wait node that was waiting has lost its run.

Forms: the log says that no users are configured, or the user has neither Password nor PasswordHash. User names ignore case, passwords don't. After 10 attempts per minute from one address the app answers 429; wait a minute.
OIDC, the app doesn't start: Pipe:Auth:Oidc needs Authority and ClientId.
OIDC, an error at the identity provider about the redirect URI: register https://<app>/signin-oidc and https://<app>/signout-callback-oidc at the client. Behind a reverse proxy see Operations.
OIDC, a certificate error in the log: the app must trust the certificate of the provider. A self-signed test certificate must be imported into the trusted root certificates.
No access although the user has the role: the role must reach the app in the claim RoleClaimType (default role). Request the scope role and check that the provider puts the roles into the token or userinfo.

The application log (console, service log or docker logs).
The execution history of the workflow: each node shows its input, output and log.