X-Git-Url: https://git.openstreetmap.org./chef.git/blobdiff_plain/3b212e01c17597264ff6303498329ffa52157f38..7b58842a8645dd4fc7cfdfe6bdffaaa9d9e8a2e3:/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 b9ffa9f44..b49a274af 100644
--- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb
+++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
@@ -1,13 +1,36 @@
 # DO NOT EDIT - This file is being maintained by Chef
 
 groups:
-  - name: alertmanager
+  - name: amsterdam
     rules:
-      - alert: prometheus target missing
-        expr: up == 0
+      - alert: pdu current draw
+        expr: rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10 > 10
         for: 5m
         labels:
-          alertgroup: "prometheus"
+          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
@@ -22,6 +45,57 @@ 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]) > 0.6
+        for: 15m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          pressure: "{{ $value | humanizePercentage }}"
   - name: database
     rules:
       - alert: postgres replication delay
@@ -31,8 +105,62 @@ groups:
           alertgroup: database
         annotations:
           delay: "{{ $value | humanizeDuration }}"
+  - name: dublin
+    rules:
+      - alert: pdu current draw
+        expr: rPDU2PhaseStatusCurrent{site="dublin",rPDU2PhaseStatusIndex="1"} / 10 > 13
+        for: 5m
+        labels:
+          alertgroup: "dublin"
+        annotations:
+          current: "{{ $value | humanize }}A"
+      - alert: site current draw
+        expr: sum(rPDU2PhaseStatusCurrent{site="dublin",rPDU2PhaseStatusIndex="1"} / 10) > 17
+        for: 5m
+        labels:
+          alertgroup: "dublin"
+        annotations:
+          current: "{{ $value | humanize }}A"
+      - alert: site temperature
+        expr: min(rPDU2SensorTempHumidityStatusTempC{site="dublin"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="dublin"}) / 10 > 25
+        for: 5m
+        labels:
+          alertgroup: "dublin"
+        annotations:
+          temperature: "{{ $value | humanize }}C"
+      - alert: site humidity
+        expr: max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="dublin"}) / 100 < 0.25 or max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="dublin"}) / 100 > 0.65
+        for: 5m
+        labels:
+          alertgroup: "dublin"
+        annotations:
+          humidity: "{{ $value | humanizePercentage }}"
+  - 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: count(fastly_healthcheck_status == 0) > 0
+        for: 15m
+        labels:
+          alertgroup: fastly
+      - alert: multipe fastly healthchecks failing
+        expr: count(fastly_healthcheck_status == 0) > 4
+        for: 5m
+        labels:
+          alertgroup: fastly
   - name: filesystem
     rules:
+      - alert: readonly filesystem
+        expr: node_filesystem_readonly == 1
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
       - alert: filesystem low on space
         expr: node_filesystem_avail_bytes / node_filesystem_size_bytes < 0.05
         for: 5m
@@ -79,6 +207,15 @@ groups:
           in_volts: "{{ with printf \"node_hwmon_in_volts{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}V{{end}}"
           in_min_volts: "{{ with printf \"node_hwmon_in_min_volts{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}V{{end}}"
           in_max_volts: "{{ with printf \"node_hwmon_in_max_volts{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}V{{end}}"
+  - name: io
+    rules:
+      - alert: io pressure
+        expr: rate(node_pressure_io_waiting_seconds_total[5m]) > 0.6
+        for: 60m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          pressure: "{{ $value | humanizePercentage }}"
   - name: ipmi
     rules:
       - alert: ipmi fan alarm
@@ -107,6 +244,34 @@ groups:
         for: 5m
         labels:
           alertgroup: "{{ $labels.instance }}"
+  - name: juniper
+    rules:
+      - alert: juniper fan alarm
+        expr: sum_over_time(jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!="running"}[5m]) > 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.site }}"
+      - alert: juniper power alarm
+        expr: sum_over_time(jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!="running"}[5m]) > 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
@@ -119,6 +284,16 @@ groups:
           active: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='active'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
           failed: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='failed'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
           spare: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='spare'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
+      - alert: mdadm array degraded
+        expr: sum (node_md_disks{state="active"}) without (state) < node_md_disks_required
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          required: "{{ with printf \"node_md_disks_required{instance='%s',device='%s'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
+          active: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='active'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
+          failed: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='failed'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
+          spare: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='spare'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
       - alert: mdadm disk failed
         expr: node_md_disks{state="failed"} > 0
         for: 0m
@@ -133,18 +308,18 @@ groups:
     rules:
       - alert: low memory
         expr: node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes < 0.1
-        for: 5m
+        for: 15m
         labels:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           memory_free: "{{ $value | humanizePercentage }}"
       - alert: memory pressure
-        expr: rate(node_vmstat_pgmajfault[1m]) > 1000
-        for: 5m
+        expr: rate(node_pressure_memory_waiting_seconds_total[5m]) > 0.6
+        for: 60m
         labels:
           alertgroup: "{{ $labels.instance }}"
         annotations:
-          major_page_faults: "{{ $value }} faults/s"
+          pressure: "{{ $value | humanizePercentage }}"
       - alert: oom kill detected
         expr: increase(node_vmstat_oom_kill[1m]) > 0
         for: 0m
@@ -169,12 +344,19 @@ 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: wireguard 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
@@ -189,6 +371,75 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           entries_used: "{{ $value | humanizePercentage }}"
+  - name: nominatim
+    rules:
+      - alert: nominatim replication delay
+        expr: nominatim_replication_delay > 10800
+        for: 1h
+        labels:
+          alertgroup: nominatim
+        annotations:
+          delay: "{{ $value | humanizeDuration }}"
+  - name: overpass
+    rules:
+      - alert: overpass osm database age
+        expr: overpass_database_age_seconds{database="osm"} > 300
+        for: 5m
+        labels:
+          alertgroup: overpass
+        annotations:
+          age: "{{ $value | humanizeDuration }}"
+      - alert: overpass area database age
+        expr: overpass_database_age_seconds{database="area"} > 86400
+        for: 1h
+        labels:
+          alertgroup: overpass
+        annotations:
+          age: "{{ $value | humanizeDuration }}"
+  - name: planet
+    rules:
+      - alert: planet dump overdue
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/(pbf|planet)/.*"} > 7 * 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 24h
+        labels:
+          alertgroup: planet
+        annotations:
+          overdue_by: "{{ $value | humanizeDuration }}"
+      - alert: notes dump overdue
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/notes/.*"} > 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 6h
+        labels:
+          alertgroup: planet
+        annotations:
+          overdue_by: "{{ $value | humanizeDuration }}"
+      - alert: daily replication feed delayed
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/day/.*"} > 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 3h
+        labels:
+          alertgroup: planet
+        annotations:
+          delayed_by: "{{ $value | humanizeDuration }}"
+      - alert: hourly replication feed delayed
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/hour/.*"} > 3600 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 30m
+        labels:
+          alertgroup: planet
+        annotations:
+          delayed_by: "{{ $value | humanizeDuration }}"
+      - alert: minutely replication feed delayed
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/minute/.*"} > 60 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 5m
+        labels:
+          alertgroup: planet
+        annotations:
+          delayed_by: "{{ $value | humanizeDuration }}"
+      - alert: changeset replication feed delayed
+        expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/changesets/.*"} > 60 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
+        for: 5m
+        labels:
+          alertgroup: planet
+        annotations:
+          delayed_by: "{{ $value | humanizeDuration }}"
   - name: postgresql
     rules:
       - alert: postgresql down
@@ -211,7 +462,7 @@ groups:
         annotations:
           connections_used: "{{ $value | humanizePercentage }}"
       - alert: postgresql deadlocks
-        expr: increase(pg_stat_database_deadlocks[1m]) > 5
+        expr: increase(pg_stat_database_deadlocks{datname!="nominatim"}[1m]) > 5
         for: 0m
         labels:
           alertgroup: "{{ $labels.instance }}"
@@ -224,6 +475,46 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           queries: "{{ $value }}"
+  - name: prometheus
+    rules:
+      - alert: prometheus configuration error
+        expr: prometheus_config_last_reload_successful == 0
+        for: 10m
+        labels:
+          alertgroup: "prometheus"
+      - alert: prometheus target missing
+        expr: up == 0
+        for: 10m
+        labels:
+          alertgroup: "prometheus"
+  - 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: rasdaemon
+    rules:
+      - alert: memory controller errors
+        expr: increase(rasdaemon_mc_events_total[1m]) > 0
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          new_errors: "{{ $value }}"
+      - alert: pcie aer errors
+        expr: increase(rasdaemon_aer_events_total[1m]) > 0
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          new_ercrors: "{{ $value }}"
   - name: smart
     rules:
       - alert: smart failure
@@ -232,7 +523,7 @@ groups:
         labels:
           alertgroup: "{{ $labels.instance }}"
       - alert: smart ssd wearout approaching
-        expr: smart_percentage_used >= 90
+        expr: smart_percentage_used >= 80
         for: 60m
         labels:
           alertgroup: "{{ $labels.instance }}"
@@ -262,18 +553,30 @@ 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 chef client 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:
@@ -299,3 +602,19 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           skew: "{{ with printf \"node_timex_offset_seconds{instance='%s'}\" $labels.instance | query }} {{ . | humanizeDuration }}{{ end }}"
+  - 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
+        for: 5m
+        labels:
+          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 }}"