What an UNREACHABLE server health status means

When VICIfast health checks stop getting any response from a Single tenant server — no HTTPS reply, no SIP (Session Initiation Protocol) registration check response, nothing at all — the server's healthStatus moves to UNREACHABLE. This is a different problem from DEGRADED, where the box still answers but reports an internal failure. UNREACHABLE means the health check system cannot contact the box. Active Agent sessions drop. Calls cannot be placed. The dialer is offline from every operational perspective until contact is restored.

Common causes of UNREACHABLE

Several different failure modes all produce the same UNREACHABLE state because they share the same outcome — the health check sends a probe and receives nothing back:

• The host is powered off or has crashed — a kernel panic, an out-of-memory kill of critical processes, or a hardware fault at the VPS provider level will all make the box unresponsive in the same way from the outside.
• Network is severed — the upstream network path to the server is blocked, a firewall rule was changed to block inbound traffic, or the server's network interface came up without an IP address after a reboot.
• Critical services failed to start after a reboot — SSH and the web server may both be down, making the server unreachable even though the hardware is powered on and the OS is running at the console.
• A full disk that caused a kernel-level write error — in severe disk-full cases the OS itself can become non-responsive because it cannot write to any filesystem, including the one the kernel uses internally.

How UNREACHABLE differs from DEGRADED in practice

The difference matters for how you respond and how quickly. A DEGRADED server is still reachable over the network — you can SSH in, look at running processes, read logs, and fix the failing service while the box continues operating. An UNREACHABLE server cannot be reached over the standard network path, so your first task is restoring connectivity before you can do anything else. That might mean using the provider's out-of-band console, checking whether the server was accidentally powered down, or reviewing firewall changes made in the last few hours.

stateDiagram-v2
  [*] --> UNKNOWN : fresh provision
  UNKNOWN --> HEALTHY : checks pass
  HEALTHY --> DEGRADED : box responds but check fails
  HEALTHY --> UNREACHABLE : no response at all
  DEGRADED --> UNREACHABLE : box stops responding entirely
  UNREACHABLE --> DEGRADED : contact restored but checks still fail
  UNREACHABLE --> HEALTHY : full recovery confirmed
  DEGRADED --> HEALTHY : streak of clean consecutive checks

Why a streak of checks is required before the badge flips

A single timeout from a health check does not immediately flip the status to UNREACHABLE. The system requires a consecutive streak of failing checks before committing to the state change. This design prevents a momentary network glitch — a brief packet loss event at the provider, a DNS hiccup, a router restart on the upstream path — from generating a false UNREACHABLE alert. By the time the badge appears, the problem has been sustained long enough to be real and warrants immediate attention.

The same streak logic applies in reverse when recovering. A single passing check after a period of UNREACHABLE does not immediately move the status back to HEALTHY. The system waits for a run of clean checks to confirm the recovery is real before updating the badge.

What to do when you see UNREACHABLE

Your first action is to try an SSH connection directly to the server's IP address. If SSH does not connect within a few seconds, the problem is at the network or OS level, not the application layer. The next step is to use the provider's out-of-band emergency console — Hetzner's rescue console, for example, lets you access the server's virtual serial port without a network connection. From there you can see whether the OS is running, check the network configuration, review kernel logs for a crash, and look for an OOM event that may have taken down critical services.

On VICIfast, an UNREACHABLE status triggers an alert and the support team can access the server's out-of-band console directly. If you are managing your own Managed hosting arrangement or a self-hosted VICIdial install, you need access to the VPS provider's emergency console before you run into this situation — not after. Setting that access up in advance is part of a solid operational runbook.

For context on how UNREACHABLE fits into the full health model alongside DEGRADED, HEALTHY, and UNKNOWN, see our server health and capacity guide. To understand the difference between health status and billing status — an important distinction when troubleshooting — read the post on server health status versus billing status.

When a server going UNREACHABLE means a call center going dark, having Managed hosting with built-in monitoring and out-of-band access is not a luxury. See VICIfast plans and get your server live in under 40 seconds.