Know before your users do
Archynt watches host CPU/RAM, service error rates and container lifecycle, and raises an alert automatically when something looks wrong — no manual thresholds to configure.
Container "fe-web-admin" stopped unexpectedly
2 minutes ago
Host CPU jumped to 91% (was ~38% over the last 30 min)
18 minutes ago
Host RAM usage at 93% — approaching exhaustion
41 minutes ago
How alerts work
CPU spike
A host's CPU jumps well above its own recent baseline within a short window — the classic "something changed and nobody noticed" case.
CPU / RAM trending up
Usage compared against the same time of day over past weeks, so a slow leak or gradual load increase gets caught before it becomes an outage.
RAM pressure
A host's memory usage gets close enough to exhaustion that an OOM kill is a real risk.
Error rate anomaly
A service's error rate breaks from its own normal baseline — compared to itself, not a fixed global threshold, so naturally noisier services don't spam.
Container stopped
A Docker container Archynt actually observes stops unexpectedly. External dependencies you don't run — a remote database, a managed service — never raise this alert.
Email + in-app notice
Every organization member gets an email the moment an alert is raised, plus a badge and toast in the dashboard for anyone who doesn't check their inbox.
Acknowledge and close
Mark an alert as acknowledged, then close it with a short comment describing the resolution. Closing is what allows the same condition to raise again later.