You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
urgent and default alert to the same webhook, and nothing routes on category=security #66
urgent and default post to the same file, so they resolve to the same
destination. The routing tree differentiates timing — group_wait: 0s and repeat_interval: 30m for critical power, against 30s/12h for the default
route — but not where the alert lands. A UPS on battery and a warning about
scrape latency arrive in the same place, differing only in how often they
repeat.
That is defensible for a one-person lab, but it is not what the two receiver
names imply, and it means there is no second channel if the first is not being
read. docs/observability.md should say which it is.
Separately: nothing routes on category
There are 10 rules labelled category: security across security.rules.yaml and network.rules.yaml — SshBruteForce, TerminalSegmentReachedInternalNetwork, IoTAttemptedLateralMovement, SuricataHighPriorityAlert and others. Every one of them routes purely on severity, so a warning-severity security alert lands in the default channel
with a 12-hour repeat interval, indistinguishable from a disk filling up.
category: power already has its own route, which is the precedent.
Activity