Monitoring and alerts

A backup that silently stopped three months ago is worse than no backup, because you believed in it. Nyx Backup reports what happened in three ways.

Desktop notifications

Each run finishes with a notification: what was backed up, how much, and any files that could not be read. Failures say what went wrong in plain language.

Notifications come from the background service, which on every platform has to hand the message to your logged-in desktop session to display it - so they work whether or not the app window is open.

Missed-backup alerts

Set a threshold per backup set and Nyx tells you when a set has not completed successfully within that many hours. This catches the failure mode that silent backups are famous for: not an error, but nothing at all - a laptop that never meets its schedule, a source disk that stopped being plugged in, a set accidentally disabled.

External dead man’s switch

The alerts above assume the machine is working well enough to complain. If it is switched off, stolen, or has lost its network, nothing raises a hand.

For that, a backup set can ping a monitoring URL on each run - healthchecks.io, Cronitor, or anything self-hosted that speaks the same convention. The monitor expects a ping on a schedule and alerts you when one does not arrive, from outside the machine entirely.

Choose which events ping:

The ping URL is not a secret in the usual sense - it grants no access to your data or your storage. Someone who obtained it could only suppress an alarm by sending a fake success, so prefer a distinct random URL per set.

Not yet available

Email and SMS alerts are not implemented. Where you need one today, the dead man’s switch above will do it - the monitoring services offer email, SMS and chat - or a failure hook script can send whatever you like. See Hook scripts.

Checking the backups themselves

Notifications tell you a run happened. Whether the stored data is still good is a separate question, answered by the integrity checks: a free, HEAD-only audit that confirms every pack is present and unchanged in size, and a deeper sampled check that downloads a percentage of chunks, decrypts them, and verifies their hashes. Both can run on a schedule, and both can be started by hand from Maintenance.