Add alerts for degraded raid arrays and failed raid disks

[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 620b44f545e720df6df7a0dcf17c36beb032a528..23bb157e475be3098725eeb230a4cde2b759d1de 100644 (file)
--- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb
+++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
@@ -5,9 +5,39 @@ groups:
     rules:
       - alert: prometheus target missing
         expr: up == 0
-        for: 5m
+        for: 10m
         labels:
           alertgroup: "prometheus"
+  - name: amsterdam
+    rules:
+      - alert: pdu current draw
+        expr: rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10 > 10
+        for: 5m
+        labels:
+          alertgroup: "amsterdam"
+        annotations:
+          current: "{{ $value | humanize }}A"
+      - alert: site current draw
+        expr: sum(rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10) > 13
+        for: 5m
+        labels:
+          alertgroup: "amsterdam"
+        annotations:
+          current: "{{ $value | humanize }}A"
+      - alert: site temperature
+        expr: min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 > 25
+        for: 5m
+        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
+        for: 5m
+        labels:
+          alertgroup: "amsterdam"
+        annotations:
+          humidity: "{{ $value | humanizePercentage }}"
   - name: apache
     rules:
       - alert: apache down
@@ -22,10 +52,52 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           busy_workers: "{{ $value | humanizePercentage }}"
+      - alert: apache low request rate
+        expr: rate(apache_accesses_total[5m]) / rate(apache_accesses_total[1h] offset 1w) < 0.25 and rate(apache_accesses_total[1h] offset 1w) > 2
+        for: 15m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          request_rate: "{{ $value | humanizePercentage }}"
+  - name: chef
+    rules:
+      - alert: chef client not running
+        expr: time() - node_systemd_timer_last_trigger_seconds{name="chef-client.timer"} > 3600
+        for: 12h
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          down_time: "{{ $value | humanizeDuration }}"
+  - name: cisco
+    rules:
+      - alert: cisco fan alarm
+        expr: rlPhdUnitEnvParamFan1Status{rlPhdUnitEnvParamFan1Status!="normal"} > 0 or rlPhdUnitEnvParamFan2Status{rlPhdUnitEnvParamFan2Status!="normal"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+        annotations:
+          fan_rpm: "{{ with printf \"rlPhdUnitEnvParamFan1Speed{site='%s',instance='%s',rlPhdUnitEnvParamStackUnit='%s'}\" $labels.site $labels.instance $labels.rlPhdUnitEnvParamStackUnit | query }}{{ . | first | value | humanize }}rpm{{end}}"
+      - alert: cisco temperature alarm
+        expr: rlPhdUnitEnvParamTempSensorStatus{rlPhdUnitEnvParamTempSensorStatus!="ok"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+        annotations:
+          temp_celsius: "{{ with printf \"rlPhdUnitEnvParamTempSensorValue{site='%s',instance='%s',rlPhdUnitEnvParamStackUnit='%s'}\" $labels.site $labels.instance $labels.rlPhdUnitEnvParamStackUnit | query }}{{ . | first | value | humanize }}C{{end}}"
+      - alert: cisco main power alarm
+        expr: rlPhdUnitEnvParamMainPSStatus{rlPhdUnitEnvParamMainPSStatus!="normal"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+      - alert: cisco redundant power alarm
+        expr: rlPhdUnitEnvParamRedundantPSStatus{rlPhdUnitEnvParamRedundantPSStatus!="normal"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
   - name: cpu
     rules:
       - alert: cpu pressure
-        expr: rate(node_pressure_cpu_waiting_seconds_total[5m]) > 0.3
+        expr: rate(node_pressure_cpu_waiting_seconds_total[5m]) > 0.6
         for: 15m
         labels:
           alertgroup: "{{ $labels.instance }}"
@@ -40,6 +112,20 @@ groups:
           alertgroup: database
         annotations:
           delay: "{{ $value | humanizeDuration }}"
+  - name: fastly
+    rules:
+      - alert: fastly error rate
+        expr: sum(rate(fastly_rt_status_group_total{status_group="5xx"}[5m])) by (service_name, datacenter) / sum(rate(fastly_rt_status_group_total[5m])) by (service_name, datacenter) > 0.005
+        for: 15m
+        labels:
+          alertgroup: fastly
+        annotations:
+          error_rate: "{{ $value | humanizePercentage }}"
+      - alert: fastly healthcheck failing
+        expr: fastly_healthcheck_status = 0
+        for: 5m
+        labels:
+          alertgroup: fastly
   - name: filesystem
     rules:
       - alert: readonly filesystem
@@ -130,6 +216,34 @@ groups:
         for: 5m
         labels:
           alertgroup: "{{ $labels.instance }}"
+  - name: juniper
+    rules:
+      - alert: juniper fan alarm
+        expr: jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!="running"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+      - alert: juniper power alarm
+        expr: jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!="running"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+  - name: mail
+    rules:
+      - alert: exim queue length
+        expr: exim_queue > exim_queue_limit
+        for: 60m
+        labels:
+          alertgroup: mail
+        annotations:
+          queue_length: "{{ $value }}"
+      - alert: mailman queue length
+        expr: mailman_queue_length > 200
+        for: 60m
+        labels:
+          alertgroup: mail
+        annotations:
+          queue_length: "{{ $value }}"
   - name: mdadm
     rules:
       - alert: mdadm array inactive
@@ -202,12 +316,19 @@ groups:
         annotations:
           bandwidth_used: "{{ $value | humanizePercentage }}"
       - alert: interface transmit errors
-        expr: rate(node_network_transmit_errs_total[1m]) / rate(node_network_transmit_packets_total[1m]) > 0.01
+        expr: rate(node_network_transmit_errs_total{device!~"wg.*"}[1m]) / rate(node_network_transmit_packets_total{device!~"wg.*"}[1m]) > 0.01
         for: 5m
         labels:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           error_rate: "{{ $value | humanizePercentage }}"
+      - alert: interface transmit errors
+        expr: rate(node_network_transmit_errs_total{device=~"wg.*"}[1m]) / rate(node_network_transmit_packets_total{device=~"wg.*"}[1m]) > 0.05
+        for: 1h
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          error_rate: "{{ $value | humanizePercentage }}"
       - alert: interface receive errors
         expr: rate(node_network_receive_errs_total[1m]) / rate(node_network_receive_packets_total[1m]) > 0.01
         for: 5m
@@ -301,6 +422,18 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           queries: "{{ $value }}"
+  - name: raid
+    rules:
+      - alert: raid array degraded
+        expr: ohai_array_info{status="degraded"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+      - alert: raid disk failed
+        expr: ohai_disk_info{status="failed"} > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
   - name: smart
     rules:
       - alert: smart failure
@@ -339,18 +472,30 @@ groups:
         for: 0m
         labels:
           alertgroup: ssl
+  - name: statuscake
+    rules:
+      - alert: statuscake uptime check failing
+        expr: statuscake_uptime{status="down",paused="false"} > 0
+        for: 10m
+        labels:
+          alertgroup: statuscake
   - name: systemd
     rules:
       - alert: systemd failed service
-        expr: node_systemd_unit_state{state="failed"} == 1
+        expr: node_systemd_unit_state{state="failed",name!="chef-client.service"} == 1
         for: 5m
         labels:
           alertgroup: "{{ $labels.instance }}"
+      - alert: systemd failed service
+        expr: node_systemd_unit_state{state="failed",name="chef-client.service"} == 1
+        for: 6h
+        labels:
+          alertgroup: "{{ $labels.instance }}"
   - name: tile
     rules:
       - alert: renderd replication delay
         expr: renderd_replication_delay > 120
-        for: 5m
+        for: 15m
         labels:
           alertgroup: tile
         annotations:
@@ -385,3 +530,10 @@ groups:
           alertgroup: web
         annotations:
           error_rate: "{{ $value | humanizePercentage }}"
+      - alert: job processing rate
+        expr: rate(pg_stat_user_tables_n_tup_del{datname="openstreetmap",relname="delayed_jobs"}[5m]) / rate(pg_stat_user_tables_n_tup_ins{datname="openstreetmap",relname="delayed_jobs"}[5m]) < 0.9 and ignoring(job, name, datname, relname, schemaname, server) chef_role{name="db-master"} == 1
+        for: 15m
+        labels:
+          alertgroup: web
+        annotations:
+          job_processing_rate: "{{ $value | humanizePercentage }}"