feat: alerts

2025-01-06 08:45:05 +01:00 · 2025-01-06 08:45:05 +01:00 · f7d73a0df5
commit f7d73a0df5
parent 8e94b34e63
10 changed files with 795 additions and 499 deletions
--- a/conf/prometheus/alerts/blackbox-exporter.yml
+++ b/conf/prometheus/alerts/blackbox-exporter.yml
@ -0,0 +1,86 @@
+groups:
+
+- name: BlackboxExporter
+
+  rules:
+
+    - alert: BlackboxProbeFailed
+      expr: 'probe_success == 0'
+      for: 0m
+      labels:
+        severity: critical
+      annotations:
+        summary: Blackbox probe failed (instance {{ $labels.instance }})
+        description: "Probe failed\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxConfigurationReloadFailure
+      expr: 'blackbox_exporter_config_last_reload_successful != 1'
+      for: 0m
+      labels:
+        severity: warning
+      annotations:
+        summary: Blackbox configuration reload failure (instance {{ $labels.instance }})
+        description: "Blackbox configuration reload failure\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxSlowProbe
+      expr: 'avg_over_time(probe_duration_seconds[1m]) > 1'
+      for: 1m
+      labels:
+        severity: warning
+      annotations:
+        summary: Blackbox slow probe (instance {{ $labels.instance }})
+        description: "Blackbox probe took more than 1s to complete\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxProbeHttpFailure
+      expr: 'probe_http_status_code <= 199 OR probe_http_status_code >= 400'
+      for: 0m
+      labels:
+        severity: critical
+      annotations:
+        summary: Blackbox probe HTTP failure (instance {{ $labels.instance }})
+        description: "HTTP status code is not 200-399\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxSslCertificateWillExpireSoon
+      expr: '3 <= round((last_over_time(probe_ssl_earliest_cert_expiry[10m]) - time()) / 86400, 0.1) < 20'
+      for: 0m
+      labels:
+        severity: warning
+      annotations:
+        summary: Blackbox SSL certificate will expire soon (instance {{ $labels.instance }})
+        description: "SSL certificate expires in less than 20 days\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxSslCertificateWillExpireSoon
+      expr: '0 <= round((last_over_time(probe_ssl_earliest_cert_expiry[10m]) - time()) / 86400, 0.1) < 3'
+      for: 0m
+      labels:
+        severity: critical
+      annotations:
+        summary: Blackbox SSL certificate will expire soon (instance {{ $labels.instance }})
+        description: "SSL certificate expires in less than 3 days\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxSslCertificateExpired
+      expr: 'round((last_over_time(probe_ssl_earliest_cert_expiry[10m]) - time()) / 86400, 0.1) < 0'
+      for: 0m
+      labels:
+        severity: critical
+      annotations:
+        summary: Blackbox SSL certificate expired (instance {{ $labels.instance }})
+        description: "SSL certificate has expired already\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxProbeSlowHttp
+      expr: 'avg_over_time(probe_http_duration_seconds[1m]) > 1'
+      for: 1m
+      labels:
+        severity: warning
+      annotations:
+        summary: Blackbox probe slow HTTP (instance {{ $labels.instance }})
+        description: "HTTP request took more than 1s\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"
+
+    - alert: BlackboxProbeSlowPing
+      expr: 'avg_over_time(probe_icmp_duration_seconds[1m]) > 1'
+      for: 1m
+      labels:
+        severity: warning
+      annotations:
+        summary: Blackbox probe slow ping (instance {{ $labels.instance }})
+        description: "Blackbox ping took more than 1s\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"