bug: unexpected daemon process

[WilliamBergamin]

Description

When using the Slack CLI to run a bolt for python project, the CLI process will sometimes terminate without cleaning up the Bolt Python process. This leaves a daemon type python process running on my machine.

This can be a pain point, I've accidentally created multiple instances of socket mode apps leaving the Slack to dispatch events randomly between them.

Version

Using lack v3.3.0

OS Info

ProductName: macOS
ProductVersion: 15.5
BuildVersion: 24F74
Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:49 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T6000

Steps to reproduce:

1. slack create my-test-app -t https://github.com/slack-samples/bolt-python-starter-template
2. slack run
3. Once the app is running, quickly disconnect and reconnect your machine from the internet
4. The Slack CLI process will terminate with the following error

🚫 Failed to stream the most recent activity logs (streaming_activity_logs_error)
   HTTP request failed (http_request_failed)

5. The python process should reconnect the socket connection
6. Triggering an event to be dispatch to your app like loading the app_home will log messages in the terminal
7. You can use the activity monitor to clean up the demon process

Expected result:

When the Slack CLI process terminates I expect that child processes it created are cleaned up as well

Actual result:

Its seems like some child processes created by the CLI are not clean up properly

Note

This may also be an issue with the hooks implementations

Requirements

• I've read and understood the Contributing guidelines and have done my best effort to follow them.
• I've read and agree to the Code of Conduct.
• I've searched for any related issues and avoided creating a duplicate issue.

[mwbrooks]

Thanks @WilliamBergamin!

One area to investigate is whether the CLI properly cancels the context when the main process is terminated unexpectedly. Cancelling the context should cause the spawned process to be terminated before the main process exits.

[zimeg]

👋 @WilliamBergamin Thanks for sharing! One workaround to stop the daemons for now can find the process IDs to kill these:

$ ps -e | grep slack-cli  # Node
$ ps -e | grep slack_cli  # Python
...
67556 ttys037    0:01.82 /path/to/project/.venv/bin/python3 -X dev -m slack_cli_hooks.hooks.start

$ kill 67556

I haven't noticed this with deno so I think this might be caused from the "SDK managed connection" handling 🤔

[zimeg]

Cancelling the context should cause the spawned process to be terminated before the main process exits.

@mwbrooks This seems like a legit problem that might be related to terminating the CLI process with a successful exit code instead of a subprocesses exiting! 🔍

The above PR #132 has attempt to avoid this error altogether and a note on this bug as it's related to the run command and apps.activities.list method were made, but I think the canceled context on all exits might be saved for a follow up if that looks good?

[zimeg]

📣 Closing this with the change to ongoing activity streaming instead of erroring in #132, which should avoid the daemon 👾

[WilliamBergamin]

Is there any change the deamon process might still be happening, for processes that stay open for a long time? 🤔

In some situations I still get them but I'm not able to reproduce it consistently yet

[zimeg]

@WilliamBergamin No doubt these are still possible! I'm hoping the errors related to apps.activities.list have been removed so far, but other cases might exist.

We might not exit long-running hook process on interrupt or unrelated exits if you are still finding this, which might be useful to investigate with more recent changes to context 🤖

I now have hopes of leaving a few experimental apps running over the weekend and looking into this more since hidden daemons in a Socket Mode setup isn't good for development.