Cannot Connect to SSH Targets (Timeout)

Boundary
1

I’m experiencing a recurring issue with Boundary where I cannot connect to SSH targets through it after some time. The issue occurs approximately once a month. When this happens:

  • SSH connections through Boundary result in a timeout.

  • Direct SSH connections (bypassing Boundary) work fine.

  • Restarting the boundary service immediately fixes the issue—SSH connections through Boundary start working again.

  • There are no errors in logs, even at trace level—only normal connection and disconnection messages.

  • Ports are open, and Boundary is listening properly (verified with netstat).

  • Worker is running within the same process as the controller on the same machine.

  • The issue is not due to too many sessions—I typically have only 1-2 active SSH sessions at a time.

Boundary version: 0.16.1

What I Have Tried:

  • Checking boundary sessions list before and after failure—sessions remain pending.

  • Running nc -vz <target_ip> 22 from the Boundary server—SSH is reachable directly.

  • Checking logs (journalctl -u boundary --since “30 min ago”)—nothing unusual, no errors.

  • Ensuring ports (9200, 9201, 9202) are open and listening.

How can I fix this issue?

1 Like