Hey guys, same here. That’s pretty much why we moved to agent-based management.
Based on my experience, once you cross a few hundred devices, SSH just doesn’t feel built for day-to-day fleet management. With an agent installed, the devices maintain an outbound connection to the management server, so we stopped worrying about VPN access or static IPs altogether.
What made the biggest difference for us was centralized control. Access is tied to admin roles and identity permissions instead of juggling SSH keys on every single box. If someone’s role changes, we update it once in the console and we’re done.
Also, visibility is way better. When we push an action to a few hundred Linux systems, we can immediately see what succeeded, what failed, and what’s just offline. No digging through playbook logs trying to piece things together.
SSH still has its place for quick troubleshooting sessions. But for actually managing a growing fleet, agent-based just feels cleaner and much easier to scale.