Set a minimum threshold for error rate alerts

[chef.git] / cookbooks / prometheus / templates / default / alert_rules.yml.erb

diff --git a/cookbooks/prometheus/templates/default/alert_rules.yml.erb b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
index 598179d5cb4a6f5303d64f601766a61b5f65e61f..b78bbaddd51fb8a54cc2834c4815f28a45c85af4 100644 (file)
--- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb
+++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
@@ -25,14 +25,14 @@ groups:
         annotations:
           current: "{{ $value | humanize }}kVA"
       - alert: site temperature
         annotations:
           current: "{{ $value | humanize }}kVA"
       - alert: site temperature

-        expr: min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 > 26
+        expr: min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 < 15 or min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 > 32

         for: 6m
         labels:
           alertgroup: "amsterdam"
         annotations:
           temperature: "{{ $value | humanize }}C"
       - alert: site humidity
         for: 6m
         labels:
           alertgroup: "amsterdam"
         annotations:
           temperature: "{{ $value | humanize }}C"
       - alert: site humidity

-        expr: max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 < 0.25 or max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 > 0.65
+        expr: max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 < 0.08 or max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 > 0.8

         for: 6m
         labels:
           alertgroup: "amsterdam"
         for: 6m
         labels:
           alertgroup: "amsterdam"


@@ -105,13 +105,20 @@ groups:
           pressure: "{{ $value | humanizePercentage }}"
   - name: database
     rules:
           pressure: "{{ $value | humanizePercentage }}"
   - name: database
     rules:

-      - alert: postgres replication delay
-        expr: pg_replication_lag_seconds > 30
-        for: 15m
+      - alert: active rails queries
+        expr: sum(pg_stat_activity_count{datname="openstreetmap",usename="rails",state="active"}) by (instance) > 50 and on (instance) chef_role{name="db-master"}
+        for: 5m

         labels:
           alertgroup: database
         annotations:
         labels:
           alertgroup: database
         annotations:

-          delay: "{{ $value | humanizeDuration }}"
+          queries: "{{ $value }}"
+      - alert: active cgimap queries
+        expr: sum(pg_stat_activity_count{datname="openstreetmap",usename="cgimap",state="active"}) by (instance) > 30 and on (instance) chef_role{name="db-master"}
+        for: 5m
+        labels:
+          alertgroup: database
+        annotations:
+          queries: "{{ $value }}"

   - name: discourse
     rules:
       - alert: discourse job failure rate
   - name: discourse
     rules:
       - alert: discourse job failure rate


@@ -414,6 +421,13 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           connections_used: "{{ $value | humanizePercentage }}"
           alertgroup: "{{ $labels.instance }}"
         annotations:
           connections_used: "{{ $value | humanizePercentage }}"

+      - alert: mysql connection errors
+        expr: increase(mysql_global_status_connection_errors_total[1m]) > 0
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          error_count: "{{ $value }}"

   - name: network
     rules:
       - alert: interface redundancy lost
   - name: network
     rules:
       - alert: interface redundancy lost


@@ -474,6 +488,11 @@ groups:
           alertgroup: nominatim
         annotations:
           delay: "{{ $value | humanizeDuration }}"
           alertgroup: nominatim
         annotations:
           delay: "{{ $value | humanizeDuration }}"

+      - alert: nominatim connections
+        expr: sum(nginx_connections_writing and on (instance) chef_role{name="nominatim"}) > 2500
+        for: 15m
+        labels:
+          alertgroup: nominatim

   - name: overpass
     rules:
       - alert: overpass osm database age
   - name: overpass
     rules:
       - alert: overpass osm database age


@@ -598,6 +617,11 @@ groups:
         for: 10m
         labels:
           alertgroup: "prometheus"
         for: 10m
         labels:
           alertgroup: "prometheus"

+      - alert: node exporter text file scrape error
+        expr: node_textfile_scrape_error > 0
+        for: 10m
+        labels:
+          alertgroup: "prometheus"

   - name: raid
     rules:
       - alert: raid controller battery failed
   - name: raid
     rules:
       - alert: raid controller battery failed


@@ -774,7 +798,7 @@ groups:
   - name: web
     rules:
       - alert: web error rate
   - name: web
     rules:
       - alert: web error rate

-        expr: sum(rate(api_call_count_total{status=~"50[0-8]|5[1-9][0-9]"}[5m])) by (instance) / sum(rate(api_call_count_total[5m])) by (instance) > 0.002
+        expr: sum(rate(api_call_count_total{status=~"50[0-8]|5[1-9][0-9]"}[5m])) by (instance) / sum(rate(api_call_count_total[5m])) by (instance) > 0.002 and sum(rate(api_call_count_total{status=~"50[0-8]|5[1-9][0-9]"}[5m])) by (instance) > 0.01

         for: 5m
         labels:
           alertgroup: web
         for: 5m
         labels:
           alertgroup: web