feat: restructure Grafana configuration, migrate alert rules to new format and add VictoriaMetrics datasource

This commit is contained in:
2025-05-31 09:27:25 +02:00
parent 17a3602d3c
commit 8b5fb0861d
14 changed files with 384 additions and 166 deletions

View File

@@ -0,0 +1,62 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
name = "HostStatusAlerts";
folder = "System Alerts";
interval = "1m";
rules = [
{
uid = "host-down-alert-uid";
title = "HostDown";
condition = "C";
data = [
{
refId = "A";
datasourceUid = "vm-datasource-uid";
queryType = "prometheus";
relativeTimeRange = { from = 60; to = 0; }; # Query over the last minute
model = {
expr = ''up'';
legendFormat = "{{instance}} ({{job}})";
instant = false; # Changed from true, as relativeTimeRange is used
};
}
{ # New Expression B: Reduce Query A
refId = "B";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
{ # Modified Expression C: Math condition based on B
refId = "C";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$B == 0"; # Check if the last value from B is 0
};
}
];
for = "2m";
noDataState = "Alerting";
execErrState = "Error";
annotations = {
summary = "Host {{ $labels.instance }} is down";
description = ''Host {{ $labels.instance }} (job: {{ $labels.job }}) has been down for more than 2 minutes.'';
};
labels = {
severity = "critical";
category = "availability";
};
}
];
}
];
}