A host that stopped reporting
Looks exactly like a healthy host on most dashboards: no alerts, no red. Here it is a state of its own, with the time it was last heard from.
Infrastructure monitoring
No findings is not the same as no problems
Silent Failure watches Linux hosts and the containers on them — and says out loud when it has no data, instead of showing you a green tick it cannot back up.
curl -fsSL https://silentfailure.app/install | sudo sh -s -- --enroll-token TOKENThe failure that is quiet
A scan that never ran
An image with no vulnerability data is a blind spot, not a clean result. It is listed as one, with the reason it has no data.
The running kernel
Is not the installed one. A patched machine that has not rebooted is still running the hole, and the host page says so.
A pool it could not ask about
A ZFS pool behind a unit that hides /dev/zfs is not a healthy pool. Unknown and healthy are drawn differently, on purpose.
What you install
One command on the host, then you approve the code it shows. The agent is a single static Go binary with no runtime dependencies.
Verified before it runs
The script carries the expected sha256, and with minisign installed it also checks a signature over the checksum file.
It asks to join
Enrollment shows a code on the host. You compare the hostname and OS in the dashboard before approving it.
It reads, it does not reach in
The agent reads /proc, the package database, listening sockets and container state. It never reads the contents of your files.
Start with one host
Free covers a single host and seven days of logs. No card needed.
Start freeRead the install docs