X-Git-Url: https://git.openstreetmap.org./chef.git/blobdiff_plain/9d85f237141a96b44e915690b54c6efc2f852227..644f12460af63db5fa70baa490def90dd7374ad7:/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 3618f1461..1fac68869 100644 --- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb +++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb @@ -3,20 +3,27 @@ groups: - name: amsterdam rules: + - alert: uplink + expr: junos_interface_up{site="amsterdam",name=~"ge-[01]/2/2"} != 1 + for: 6m + labels: + alertgroup: "amsterdam" + annotations: + status: "{{ $value }}" - alert: pdu current draw - expr: rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10 > 10 + expr: rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10 > 28 for: 6m labels: alertgroup: "amsterdam" annotations: current: "{{ $value | humanize }}A" - - alert: site current draw - expr: sum(rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10) > 13 + - alert: site power + expr: sum(avg_over_time(rPDU2PhaseStatusApparentPower{site="amsterdam",rPDU2PhaseStatusIndex="1"}[1h]) / 100) > 3.5 for: 6m labels: alertgroup: "amsterdam" annotations: - current: "{{ $value | humanize }}A" + current: "{{ $value | humanize }}kVA" - alert: site temperature expr: min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 > 26 for: 6m @@ -45,13 +52,13 @@ 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 + - alert: apache connection limit + expr: (apache_connections{state="total"} - on (instance) apache_connections{state="closing"}) / on (instance) (apache_server_limit * on (instance) (apache_threads_per_child + on (instance) (apache_async_request_worker_factor * on (instance) apache_workers{state="idle"} / on(instance) apache_processes{state="all"}))) > 0.8 + for: 5m labels: alertgroup: "{{ $labels.instance }}" annotations: - request_rate: "{{ $value | humanizePercentage }}" + connections: "{{ $value | humanizePercentage }}" - name: chef rules: - alert: chef client not running @@ -90,37 +97,60 @@ groups: - name: cpu rules: - alert: cpu pressure - expr: rate(node_pressure_cpu_waiting_seconds_total[5m]) > 0.6 - for: 15m + expr: rate(node_pressure_cpu_waiting_seconds_total[5m]) > 0.75 + for: 60m labels: alertgroup: "{{ $labels.instance }}" annotations: pressure: "{{ $value | humanizePercentage }}" - name: database rules: - - alert: postgres replication delay - expr: pg_replication_lag_seconds > 5 + - 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: - 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: + delay: "{{ $value }}" + - name: discourse + rules: + - alert: discourse job failure rate + expr: rate(discourse_job_failures[5m]) > 0 + for: 5m + labels: + alertgroup: discourse + annotations: + failure_rate: "{{ $value }} jobs/s" - name: dublin rules: + - alert: uplink + expr: junos_interface_up{site="dublin",name=~"ge-[01]/2/2"} != 1 + for: 6m + labels: + alertgroup: "dublin" + annotations: + status: "{{ $value }}" - alert: pdu current draw - expr: rPDU2PhaseStatusCurrent{site="dublin",rPDU2PhaseStatusIndex="1"} / 10 > 13 + expr: rPDU2PhaseStatusCurrent{site="dublin",rPDU2PhaseStatusIndex="1"} / 10 > 28 for: 6m labels: alertgroup: "dublin" annotations: current: "{{ $value | humanize }}A" - - alert: site current draw - expr: sum(rPDU2PhaseStatusCurrent{site="dublin",rPDU2PhaseStatusIndex="1"} / 10) > 17 + - alert: site power + expr: sum(avg_over_time(rPDU2PhaseStatusApparentPower{site="dublin",rPDU2PhaseStatusIndex="1"}[1h]) / 100) > 4 for: 6m labels: alertgroup: "dublin" annotations: - current: "{{ $value | humanize }}A" + current: "{{ $value | humanize }}kVA" - alert: site temperature expr: min(rPDU2SensorTempHumidityStatusTempC{site="dublin"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="dublin"}) / 10 > 26 for: 6m @@ -144,13 +174,23 @@ groups: alertgroup: fastly annotations: error_rate: "{{ $value | humanizePercentage }}" - - alert: fastly healthcheck failing - expr: count(fastly_healthcheck_status == 0) by (service) > 0 + - alert: fastly frontend healthcheck warning + expr: count(fastly_healthcheck_status == 0) by (service, datacenter) > 2 + for: 15m + labels: + alertgroup: fastly + - alert: fastly frontend healthcheck critical + expr: count(fastly_healthcheck_status == 0) by (service, datacenter) == count(fastly_healthcheck_status) by (service, datacenter) + for: 5m + labels: + alertgroup: fastly + - alert: fastly backend healthcheck warning + expr: count(fastly_healthcheck_status == 0) by (service, backend) > 10 for: 15m labels: alertgroup: fastly - - alert: multiple fastly healthchecks failing - expr: count(fastly_healthcheck_status == 0) by (service) > 4 + - alert: fastly backend healthcheck critical + expr: count(fastly_healthcheck_status == 0) by (service, backend) == count(fastly_healthcheck_status) by (service, backend) for: 5m labels: alertgroup: fastly @@ -246,21 +286,51 @@ groups: alertgroup: "{{ $labels.instance }}" - name: juniper rules: + - alert: juniper red alarms + expr: juniper_alarms_red_count > 0 + for: 5m + labels: + alertgroup: "{{ $labels.site }}" + annotations: + alarm_count: "{{ $value }} alarms" + - alert: juniper yellow alarms + expr: juniper_alarms_yellow_count > 0 + for: 5m + labels: + alertgroup: "{{ $labels.site }}" + annotations: + alarm_count: "{{ $value }} alarms" - alert: juniper cpu alarm - expr: jnxOperatingCPU{jnxOperatingContentsIndex="7"} > 30 + expr: junos_route_engine_load_average_five / 2 > 0.5 for: 5m labels: alertgroup: "{{ $labels.site }}" + annotations: + load_average: "{{ $value | humanizePercentage }}" - alert: juniper fan alarm - expr: jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!~"running.*"} > 0 + expr: junos_environment_fan_up != 1 for: 5m labels: alertgroup: "{{ $labels.site }}" - alert: juniper power alarm - expr: jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!~"running.*"} > 0 + expr: junos_environment_power_up != 1 + for: 5m + labels: + alertgroup: "{{ $labels.site }}" + - alert: juniper laser receive power + expr: junos_interface_diagnostics_laser_rx_dbm < -12 and on (site, instance, name) junos_interface_admin_up == 1 + for: 5m + labels: + alertgroup: "{{ $labels.site }}" + annotations: + power: "{{ $value }} dBm" + - alert: juniper laser transmit power + expr: junos_interface_diagnostics_laser_output_dbm < -8 and on (site, instance, name) junos_interface_admin_up == 1 for: 5m labels: alertgroup: "{{ $labels.site }}" + annotations: + power: "{{ $value }} dBm" - name: mail rules: - alert: exim down @@ -269,7 +339,7 @@ groups: labels: alertgroup: "{{ $labels.instance }}" - alert: exim queue length - expr: exim_queue > exim_queue_limit + expr: exim_queue > ignoring(job) exim_queue_limit for: 60m labels: alertgroup: mail @@ -353,15 +423,22 @@ groups: connections_used: "{{ $value | humanizePercentage }}" - name: network rules: + - alert: interface redundancy lost + expr: node_bonding_active < 2 and on (instance, master) label_replace(chef_network_interface{bond_mode="802.3ad"}, "master", "$1", "name", "(.*)") + for: 5m + labels: + alertgroup: "{{ $labels.instance }}" + annotations: + link_count: "{{ $value }}" - alert: interface transmit rate - expr: rate(node_network_transmit_bytes_total[1m]) / node_network_speed_bytes > 0.98 + expr: rate(node_network_transmit_bytes_total[1m]) / node_network_speed_bytes > 0.99 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 + expr: rate(node_network_receive_bytes_total[1m]) / node_network_speed_bytes > 0.99 for: 5m labels: alertgroup: "{{ $labels.instance }}" @@ -489,8 +566,8 @@ groups: labels: alertgroup: "{{ $labels.instance }}" - alert: postgresql replication delay - expr: pg_replication_lag_seconds > 5 - for: 1m + expr: pg_replication_lag_seconds > 30 + for: 15m labels: alertgroup: "{{ $labels.instance }}" annotations: @@ -509,8 +586,8 @@ groups: alertgroup: "{{ $labels.instance }}" annotations: new_deadlocks: "{{ $value }}" - - alert: postgresql slow queries - expr: pg_slow_queries > 0 + - alert: postgresql idle transactions + expr: sum(pg_process_idle_seconds_count{state="idle in transaction"}) by (instance, server) > sum(pg_process_idle_seconds_bucket{state="idle in transaction",le="300"}) by (instance, server) for: 5m labels: alertgroup: "{{ $labels.instance }}" @@ -530,6 +607,16 @@ groups: alertgroup: "prometheus" - name: raid rules: + - alert: raid controller battery failed + expr: ohai_controller_info{battery_status="failed"} > 0 + for: 5m + labels: + alertgroup: "{{ $labels.instance }}" + - alert: raid controller battery recharging + expr: ohai_controller_info{battery_status="recharging"} > 0 + for: 4h + labels: + alertgroup: "{{ $labels.instance }}" - alert: raid array degraded expr: ohai_array_info{status="degraded"} > 0 for: 5m @@ -564,12 +651,21 @@ groups: labels: alertgroup: "{{ $labels.instance }}" - alert: smart ssd wearout approaching - expr: smart_percentage_used >= 80 + expr: smart_percentage_used / 100 >= 0.8 for: 60m labels: alertgroup: "{{ $labels.instance }}" annotations: percentage_used: "{{ $value | humanizePercentage }}" + - name: smokeping + rules: + - alert: packet loss + expr: 1 - (rate(smokeping_response_duration_seconds_count[5m]) / rate(smokeping_requests_total[5m])) > 0.02 + for: 10m + labels: + alertgroup: smokeping + annotations: + loss_rate: "{{ $value | humanizePercentage }}" - name: snmp rules: - alert: snmp pdus missing @@ -606,7 +702,7 @@ groups: - name: statuscake rules: - alert: statuscake uptime check failing - expr: statuscake_uptime{status="down",paused="false"} > 0 + expr: statuscake_paused == 0 and statuscake_up == 0 for: 10m labels: alertgroup: statuscake @@ -618,10 +714,33 @@ groups: labels: alertgroup: "{{ $labels.instance }}" - alert: systemd failed chef client service - expr: node_systemd_unit_state{state="failed",name="chef-client.service"} == 1 - for: 6h + expr: sum_over_time(node_systemd_unit_state{state="inactive",name="chef-client.service"}[6h]) == 0 + for: 0m labels: alertgroup: "{{ $labels.instance }}" + - name: taginfo + rules: + - alert: taginfo planet age + expr: time() - taginfo_data_from_seconds > 129600 and on (instance) chef_role{name="taginfo"} + for: 0m + labels: + alertgroup: taginfo + annotations: + age: "{{ $value | humanizeDuration }}" + - alert: taginfo database age + expr: time() - taginfo_database_update_finish_seconds > 129600 and on (instance) chef_role{name="taginfo"} + for: 0m + labels: + alertgroup: taginfo + annotations: + age: "{{ $value | humanizeDuration }}" + - alert: taginfo database size + expr: abs(delta(taginfo_database_size_bytes[30m])) / taginfo_database_size_bytes > 0.1 + for: 30m + labels: + alertgroup: taginfo + annotations: + size_change: "{{ $value | humanizePercentage }}" - name: tile rules: - alert: renderd replication delay @@ -639,8 +758,8 @@ groups: annotations: miss_rate: "{{ $value | humanizePercentage }}" - alert: tile render rate - expr: sum(rate(renderd_zoom_metatiles_total[5m])) by (instance) < 1 - for: 5m + expr: sum(rate(renderd_zoom_metatiles_total[5m])) by (instance) == 0 + for: 15m labels: alertgroup: tile annotations: @@ -669,8 +788,8 @@ groups: 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 + expr: rate(pg_stat_user_tables_n_tup_del{datname="openstreetmap",relname="delayed_jobs"}[1h]) / rate(pg_stat_user_tables_n_tup_ins{datname="openstreetmap",relname="delayed_jobs"}[1h]) < 0.9 and ignoring(job, name, datname, relname, schemaname, server) chef_role{name="db-master"} == 1 + for: 1h labels: alertgroup: web annotations: