Any way to view connected workers?

I’m trying to diagnose why I can’t connect to one of my targets. From the controller, I have no way of of seeing which workers are registered and how sessions are being routed (right? I’m not missing something here?). This makes it extremely difficult to understand connection problems in a faulty deployment.

Boundary doesn’t tell me if a connection couldn’t be established with one of my hosts. Heck, I can make a target with some random IP address that doesn’t exist, tell it to connect, and it will hand me back a session address if everything is fine (although when I stick the address in my application, no connection will be made).

Am I missing something? Is there a technique built in to diagnose connection issues? Or do I need to dump the postgres database and look at the raw data?

1 Like

Best thing I’ve found for diagnosing connection issues are the Controller logs, which have given me hints like the port number, IP & “connection refused” (or similar).

How do I access those logs exactly?

The Controller process will spit them out as JSON to standard out, by default. So if you you’re running Boundary direct from a terminal then make a failing connection and immediately check there.

If you’re running via systemd then it will likely be sudo journalctl -u {your boundary service name} -e . The -e will take you to the end of the logs, -f can be used to follow.

This is what a log from the terminal looked like for me, during a failure…

2 Likes