X-Git-Url: https://git.openstreetmap.org./chef.git/blobdiff_plain/2c385f3d70fb6867de8c6f0d846eba62fba1aeb0..44b52281d387a2a26df9cd9ac5e228c0976089fd:/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 d870444ca..23b94727c 100644
--- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb
+++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
@@ -1,13 +1,154 @@
 # 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
+        expr: apache_up == 0
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+      - alert: apache workers busy
+        expr: sum(apache_workers{state="busy"}) by (instance) / sum(apache_scoreboard) by (instance) > 0.8
+        for: 5m
+        labels:
+          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
+        expr: pg_replication_lag_seconds > 5
+        for: 5m
+        labels:
+          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: count(fastly_healthcheck_status == 0) > 0
+        for: 15m
+        labels:
+          alertgroup: fastly
+      - alert: fastly healthcheck 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
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          percentage_free: "{{ $value | humanizePercentage }}"
+          free_bytes: "{{ with printf \"node_filesystem_avail_bytes{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value | humanize1024 }}bytes{{end}}"
+          total_total: "{{ with printf \"node_filesystem_size_bytes{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value | humanize1024 }}bytes{{end}}"
+      - alert: filesystem low on inodes
+        expr: node_filesystem_files_free / node_filesystem_files < 0.1
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          percentage_free: "{{ $value | humanizePercentage }}"
+          free_inodes: "{{ with printf \"node_filesystem_files_free{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value }}{{end}}"
+          total_inodes: "{{ with printf \"node_filesystem_files{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value }}{{end}}"
   - name: hwmon
     rules:
       - alert: hwmon fan alarm
@@ -36,6 +177,15 @@ 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
@@ -59,6 +209,39 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           voltage_volts: "{{ with printf \"ipmi_voltage_volts{instance='%s',id='%s'}\" $labels.instance $labels.id | query }}{{ . | first | value | humanize }}V{{end}}"
+      - alert: ipmi power alarm
+        expr: ipmi_power_state > 0 or ipmi_sensor_state{type=~"Power .*"} > 0
+        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
@@ -71,6 +254,16 @@ 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
@@ -85,18 +278,18 @@ 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
@@ -104,3 +297,269 @@ groups:
           alertgroup: "{{ $labels.instance }}"
         annotations:
           new_oom_kills: "{{ $value }}"
+  - name: network
+    rules:
+      - alert: interface transmit rate
+        expr: rate(node_network_transmit_bytes_total[1m]) / node_network_speed_bytes > 0.98
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          bandwidth_used: "{{ $value | humanizePercentage }}"
+      - alert: interface receive rate
+        expr: rate(node_network_receive_bytes_total[1m]) / node_network_speed_bytes > 0.98
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          bandwidth_used: "{{ $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]) > 0.01
+        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
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          error_rate: "{{ $value | humanizePercentage }}"
+      - alert: conntrack entries
+        expr: node_nf_conntrack_entries / node_nf_conntrack_entries_limit > 0.8
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          entries_used: "{{ $value | humanizePercentage }}"
+  - 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
+        expr: pg_up == 0
+        for: 1m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+      - alert: postgresql replication delay
+        expr: pg_replication_lag_seconds > 5
+        for: 1m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          delay: "{{ $value | humanizeDuration }}"
+      - alert: postgresql connection limit
+        expr: sum (pg_stat_activity_count) by (instance, server) / sum (pg_settings_max_connections) by (instance, server) > 0.8
+        for: 1m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          connections_used: "{{ $value | humanizePercentage }}"
+      - alert: postgresql deadlocks
+        expr: increase(pg_stat_database_deadlocks{datname!="nominatim"}[1m]) > 5
+        for: 0m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          new_deadlocks: "{{ $value }}"
+      - alert: postgresql slow queries
+        expr: pg_slow_queries > 0
+        for: 5m
+        labels:
+          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
+        expr: smart_health_status == 0
+        for: 60m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+      - alert: smart ssd wearout approaching
+        expr: smart_percentage_used >= 80
+        for: 60m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+        annotations:
+          percentage_used: "{{ $value | humanizePercentage }}"
+  - name: ssl
+    rules:
+      - alert: ssl certificate probe failed
+        expr: ssl_probe_success == 0
+        for: 60m
+        labels:
+          alertgroup: ssl
+      - alert: ssl certificate expiry
+        expr: ssl_verified_cert_not_after{chain_no="0"} - time() < 86400 * 14
+        for: 0m
+        labels:
+          alertgroup: ssl
+        annotations:
+          expires_in: "{{ $value | humanizeDuration }}"
+      - alert: ssl certificate revoked
+        expr: ssl_ocsp_response_status == 1
+        for: 0m
+        labels:
+          alertgroup: ssl
+      - alert: ocsp status unknown
+        expr: ssl_ocsp_response_status == 1
+        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",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: 15m
+        labels:
+          alertgroup: tile
+        annotations:
+          delay: "{{ $value | humanizeDuration }}"
+      - alert: missed tile rate
+        expr: sum(rate(modtile_http_response_total{code="404"}[5m])) by (instance) / sum(rate(modtile_http_response_total[5m])) by (instance) > 0.05
+        for: 5m
+        labels:
+          alertgroup: tile
+        annotations:
+          miss_rate: "{{ $value | humanizePercentage }}"
+  - name: time
+    rules:
+      - alert: clock not synchronising
+        expr: min_over_time(node_timex_sync_status[1m]) == 0 and node_timex_maxerror_seconds >= 16
+        for: 5m
+        labels:
+          alertgroup: "{{ $labels.instance }}"
+      - alert: clock skew detected
+        expr: (node_timex_offset_seconds > 0.05 and deriv(node_timex_offset_seconds[5m]) >= 0) or (node_timex_offset_seconds < -0.05 and deriv(node_timex_offset_seconds[5m]) <= 0)
+        for: 5m
+        labels:
+          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 }}"