rafaelldi / aspire-plugin Goto Github PK

Error log from my dapr sidecars, note I'm ommitting computer/app specific values:

time="2024-01-26T11:50:06.496542-05:00" level=warning msg="failed to subscribe to topics: error occurred while beginning pubsub {{PUB_SUB_NAME}}: Get \"http://127.0.0.1:5034/dapr/subscribe\": read tcp 127.0.0.1:60618->127.0.0.1:5034: read: connection reset by peer " app_id={{APP_ID}} instance={{INSTANCE_ID}} scope=dapr.runtime type=log ver=1.12.4

This only happens when running with this plugin. I'm on version 2023.3.3 of Rider and on MacOS version 14.3. My dapr CLI version is 1.12.0 and runtime version is 1.12.4. If I run the app via dotnet run on the command line, pubsub works fine. So I'm guessing it's something within this plugin or Rider.

https://learn.microsoft.com/en-us/dotnet/aspire/whats-new/preview-3

When I make a blank Aspire project, it launches successfully. But when I try to run my upgraded Aspire-based project, it just does nothing.

When I try to run the AppHost, the projects that are set with direct path do not function properly and debugging doesn't work for them.

Example:

IResourceBuilder<ProjectResource> usersApi =
	builder.AddProject("users", "../../../Source/Services/Bargeh.Users.Api/Bargeh.Users.API.csproj")
		   .WithReference(postgres);

I have update my .NET Aspire version to Preview 5 and plugin starting AppHost with Aspire.Hosting.Dcp.dcpctrl.ExecutableReconciler iterated.

If I start the AppHost without using the plugin, it work fine.

I think Aspire have changed API for connection in latest version, and plugin can`t connect to AppHost and start work.

I would upload the logs here, but there might be sensitive environment variables in the logs.

Update

I think this is an issue with sudo on macOS and Linux. Workloads need to use sudo dotnet workload update aspire and escalate to administrator.

Hello,

I'm writing to report an issue I've encountered with the .NET Aspire plugin for Rider. The plugin was functioning perfectly on my Windows 11 machine, detecting all projects, running them, and attaching the debugger as required. However, when I attempted to open the project on a Linux PC, I ran into several errors due to the use of "\" in path strings and case sensitivity issues with paths. After updating the project files, I was able to run the project on Linux.

Upon returning to my Windows machine, pulling the changes, and running the project, I found that only one project was successful. The AppHost was built and running, but out of all the projects, only one ran. I noticed that reverting to the commits before the Linux changes resolved the issue. Interestingly, when I run the AppHost with .NET CLI, all projects are built and run successfully. I've tried to capture the build logs but found no errors.

I'm unsure why this is happening or what the exact issue is. Any assistance would be greatly appreciated.

Thank you.

I've a Preview 3, when I F5 run the solution debugging works. With Shift+F5 processes are stopping but not the AppHost.

If I make a Hot reload compatible change and hit "Apply" in rider, the given process is not reloaded, so I've to stop everything, and re-run.

They're not handled properly right now

https://opentelemetry.io/docs/specs/semconv/messaging/messaging-spans/
https://opentelemetry.io/docs/specs/semconv/attributes-registry/messaging/

https://opentelemetry.io/docs/specs/semconv/database/database-spans/
https://opentelemetry.io/docs/specs/semconv/attributes-registry/db/

VS.net works fine with the following and the path is correct:

		"environmentVariables": {
			"ASPNETCORE_ENVIRONMENT": "Development",
			"DOTNET_ENVIRONMENT": "Development",
			"DOTNET_DASHBOARD_OTLP_ENDPOINT_URL": "http://localhost:7064",
			"DEV_PATH": "$(MSBuildProjectDirectory)/../.dev",
			"CERTIFICATE_PATH": "$(MSBuildProjectDirectory)/../.dev/certificates",
			"CONFIG_PATH": "$(MSBuildProjectDirectory)/../Configuration"
		}

Rider is fine with those same types of configs in each of the individual services that are run if they're run by themselves.

But Rider with the host running, doesn't get the right pathing for the host. It appears to ignore the value of $(MSBuildProjectdirectory) which is null when dumped to the console in the host, but not if I run it from vs.net or run any of the services separately in rider.

In fact, it appears that the standard environment variables aren't set at all and don't get passed into the services launched either.

This is a blocker for us.

https://opentelemetry.io/docs/specs/semconv/attributes-registry/http/

The current mechanism is based on the run configurations. This causes some problems:

• #64
• #72

So, it would be better to rework this mechanism.

I am using the following project DaprOutbox.

When I use the Aspire Host run configuration added by the plugin to run, I see the following issues in the console logs of webfrontend-dapr and/or subscriber-dapr:

time="2024-01-05T15:21:19.0975586+01:00" level=warning msg="failed to subscribe to topics: error occurred while beginning pubsub pubsub: Get \"http://127.0.0.1:5203/dapr/subscribe\": read tcp 127.0.0.1:63905->127.0.0.1:5203: wsarecv: An existing connection was forcibly closed by the remote host. " app_id=subscriber-dapr instance=Beast scope=dapr.runtime type=log ver=1.12.2

When I run using the .NET Launch Settings Profile; DaprOutbox.AppHost: http then everything works fine but then I don't have access to all features provided by your plugin.

In latest versions, is impossible to start using plugin without normalize Environment Variables
Logs have no errors or detailed information. All looks like work fine
Additionaly need to fix Build Project to Build Solution

For plugin users, look #109 for information how to fix this problem.

I don't know if a plugin is able to do this, but it would be interesting if a database connection would also be added when debugging a test that uses the DistributedApplicationTestingBuilder. The connection string seems to be reachable through the DistributedApplicationModel.Resources list. For example, a Postgres connection string could be accessed by casting it to PostgresServerResource and reading the ConnectionString value by calling await ConnectionString.GetValueAsync(CancellationToken.None)

If I hit restart on the debug output window for a given project that is attached to the host it doesn't reconnect properly in the dashboard etc. I'm not sure if this is aspire or this plugin.

I upgraded my Rider to 2024.1, and now when I start my Aspire project using the plugin, 2 of my projects are not starting.

If I start the AppHost without using the plugin, it seems to work fine. The logs are empty in both the Aspire Dashboard but I see this in AppHost console in Rider:

fail: Aspire.Hosting.Dcp.dcpctrl.ExecutableReconciler[0]
      run session could not be started  {"Executable": {"name":"account-management-api-2vkhlu8"}, "Reconciliation": 7, "Status": "500 Internal Server Error", "Body": ""}
fail: Aspire.Hosting.Dcp.dcpctrl.ExecutableReconciler[0]
      run session could not be started  {"Executable": {"name":"app-gateway-cuu95rg"}, "Reconciliation": 12, "Status": "500 Internal Server Error", "Body": ""}

This is an open-source project, so you can clone https://github.com/platformplatform/PlatformPlatform and run the /application/PlatformPlatform.slnto reproduce.

I happened to have a screenshot from yesterday, of how it looked before the update:

Something happened in the latest version, because this is now the wrong directory compared to what it would be if you debugged the service directly. It's once deeper that it should be as thus all of the launchsettings.json files for every service are wrong and you can't debug host, and debug the service by itself without having to change the profile's pathing for environment variables.