cookbooks/prometheus/templates/default/alert_rules.yml.erb

   1 # DO NOT EDIT - This file is being maintained by Chef
   2
   3 groups:
   4   - name: alertmanager
   5     rules:
   6       - alert: prometheus target missing
   7         expr: up == 0
   8         for: 5m
   9         labels:
  10           alertgroup: "prometheus"
  11   - name: amsterdam
  12     rules:
  13       - alert: pdu current draw
  14         expr: rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10 > 10
  15         for: 5m
  16         labels:
  17           alertgroup: "amsterdam"
  18         annotations:
  19           current: "{{ $value | humanize }}A"
  20       - alert: site current draw
  21         expr: sum(rPDU2PhaseStatusCurrent{site="amsterdam",rPDU2PhaseStatusIndex="1"} / 10) > 13
  22         for: 5m
  23         labels:
  24           alertgroup: "amsterdam"
  25         annotations:
  26           current: "{{ $value | humanize }}A"
  27       - alert: site temperature
  28         expr: min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 < 18 or min(rPDU2SensorTempHumidityStatusTempC{site="amsterdam"}) / 10 > 25
  29         for: 5m
  30         labels:
  31           alertgroup: "amsterdam"
  32         annotations:
  33           temperature: "{{ $value | humanize }}C"
  34       - alert: site humidity
  35         expr: max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 < 0.25 or max(rPDU2SensorTempHumidityStatusRelativeHumidity{site="amsterdam"}) / 100 > 0.65
  36         for: 5m
  37         labels:
  38           alertgroup: "amsterdam"
  39         annotations:
  40           humidity: "{{ $value | humanizePercentage }}"
  41   - name: apache
  42     rules:
  43       - alert: apache down
  44         expr: apache_up == 0
  45         for: 5m
  46         labels:
  47           alertgroup: "{{ $labels.instance }}"
  48       - alert: apache workers busy
  49         expr: sum(apache_workers{state="busy"}) by (instance) / sum(apache_scoreboard) by (instance) > 0.8
  50         for: 5m
  51         labels:
  52           alertgroup: "{{ $labels.instance }}"
  53         annotations:
  54           busy_workers: "{{ $value | humanizePercentage }}"
  55       - alert: apache low request rate
  56         expr: rate(apache_accesses_total[5m]) / rate(apache_accesses_total[1h] offset 1w) < 0.25 and rate(apache_accesses_total[1h] offset 1w) > 2
  57         for: 15m
  58         labels:
  59           alertgroup: "{{ $labels.instance }}"
  60         annotations:
  61           request_rate: "{{ $value | humanizePercentage }}"
  62   - name: chef
  63     rules:
  64       - alert: chef client not running
  65         expr: time() - node_systemd_timer_last_trigger_seconds{name="chef-client.timer"} > 3600
  66         for: 12h
  67         labels:
  68           alertgroup: "{{ $labels.instance }}"
  69         annotations:
  70           down_time: "{{ $value | humanizeDuration }}"
  71   - name: cisco
  72     rules:
  73       - alert: cisco fan alarm
  74         expr: rlPhdUnitEnvParamFan1Status{rlPhdUnitEnvParamFan1Status!="normal"} > 0 or rlPhdUnitEnvParamFan2Status{rlPhdUnitEnvParamFan2Status!="normal"} > 0
  75         for: 5m
  76         labels:
  77           alertgroup: "{{ $labels.site }}"
  78       - alert: cisco temperature alarm
  79         expr: rlPhdUnitEnvParamTempSensorStatus{rlPhdUnitEnvParamTempSensorStatus!="ok"} > 0
  80         for: 5m
  81         labels:
  82           alertgroup: "{{ $labels.site }}"
  83       - alert: cisco main power alarm
  84         expr: rlPhdUnitEnvParamMainPSStatus{rlPhdUnitEnvParamMainPSStatus!="normal"} > 0
  85         for: 5m
  86         labels:
  87           alertgroup: "{{ $labels.site }}"
  88       - alert: cisco redundant power alarm
  89         expr: rlPhdUnitEnvParamRedundantPSStatus{rlPhdUnitEnvParamRedundantPSStatus!="normal"} > 0
  90         for: 5m
  91         labels:
  92           alertgroup: "{{ $labels.site }}"
  93   - name: cpu
  94     rules:
  95       - alert: cpu pressure
  96         expr: rate(node_pressure_cpu_waiting_seconds_total[5m]) > 0.6
  97         for: 15m
  98         labels:
  99           alertgroup: "{{ $labels.instance }}"
 100         annotations:
 101           pressure: "{{ $value | humanizePercentage }}"
 102   - name: database
 103     rules:
 104       - alert: postgres replication delay
 105         expr: pg_replication_lag_seconds > 5
 106         for: 5m
 107         labels:
 108           alertgroup: database
 109         annotations:
 110           delay: "{{ $value | humanizeDuration }}"
 111   - name: fastly
 112     rules:
 113       - alert: error rate
 114         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
 115         for: 15m
 116         labels:
 117           alertgroup: fastly
 118         annotations:
 119           error_rate: "{{ $value | humanizePercentage }}"
 120   - name: filesystem
 121     rules:
 122       - alert: readonly filesystem
 123         expr: node_filesystem_readonly == 1
 124         for: 0m
 125         labels:
 126           alertgroup: "{{ $labels.instance }}"
 127       - alert: filesystem low on space
 128         expr: node_filesystem_avail_bytes / node_filesystem_size_bytes < 0.05
 129         for: 5m
 130         labels:
 131           alertgroup: "{{ $labels.instance }}"
 132         annotations:
 133           percentage_free: "{{ $value | humanizePercentage }}"
 134           free_bytes: "{{ with printf \"node_filesystem_avail_bytes{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value | humanize1024 }}bytes{{end}}"
 135           total_total: "{{ with printf \"node_filesystem_size_bytes{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value | humanize1024 }}bytes{{end}}"
 136       - alert: filesystem low on inodes
 137         expr: node_filesystem_files_free / node_filesystem_files < 0.1
 138         for: 5m
 139         labels:
 140           alertgroup: "{{ $labels.instance }}"
 141         annotations:
 142           percentage_free: "{{ $value | humanizePercentage }}"
 143           free_inodes: "{{ with printf \"node_filesystem_files_free{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value }}{{end}}"
 144           total_inodes: "{{ with printf \"node_filesystem_files{instance='%s',mountpoint='%s'}\" $labels.instance $labels.mountpoint | query }}{{ . | first | value }}{{end}}"
 145   - name: hwmon
 146     rules:
 147       - alert: hwmon fan alarm
 148         expr: node_hwmon_fan_alarm == 1
 149         for: 5m
 150         labels:
 151           alertgroup: "{{ $labels.instance }}"
 152         annotations:
 153           fan_rpm: "{{ with printf \"node_hwmon_fan_rpm{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}rpm{{end}}"
 154           fan_min_rpm: "{{ with printf \"node_hwmon_fan_min_rpm{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}rpm{{end}}"
 155       - alert: hwmon temperature alarm
 156         expr: node_hwmon_temp_alarm == 1
 157         for: 5m
 158         labels:
 159           alertgroup: "{{ $labels.instance }}"
 160         annotations:
 161           temp_celsius: "{{ with printf \"node_hwmon_temp_celsius{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}C{{end}}"
 162           temp_max_celsius: "{{ with printf \"node_hwmon_temp_max_celsius{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}C{{end}}"
 163           temp_crit_celsius: "{{ with printf \"node_hwmon_temp_crit_celsius{instance='%s',chip='%s',sensor='%s'}\" $labels.instance $labels.chip $labels.sensor | query }}{{ . | first | value | humanize }}C{{end}}"
 164       - alert: hwmon voltage alarm
 165         expr: node_hwmon_in_alarm == 1
 166         for: 5m
 167         labels:
 168           alertgroup: "{{ $labels.instance }}"
 169         annotations:
 170           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}}"
 171           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}}"
 172           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}}"
 173   - name: io
 174     rules:
 175       - alert: io pressure
 176         expr: rate(node_pressure_io_waiting_seconds_total[5m]) > 0.6
 177         for: 60m
 178         labels:
 179           alertgroup: "{{ $labels.instance }}"
 180         annotations:
 181           pressure: "{{ $value | humanizePercentage }}"
 182   - name: ipmi
 183     rules:
 184       - alert: ipmi fan alarm
 185         expr: ipmi_fan_speed_state > 0
 186         for: 5m
 187         labels:
 188           alertgroup: "{{ $labels.instance }}"
 189         annotations:
 190           fan_speed_rpm: "{{ with printf \"ipmi_fan_speed_rpm{instance='%s',id='%s'}\" $labels.instance $labels.id | query }}{{ . | first | value | humanize }}rpm{{end}}"
 191       - alert: ipmi temperature alarm
 192         expr: ipmi_temperature_state > 0
 193         for: 5m
 194         labels:
 195           alertgroup: "{{ $labels.instance }}"
 196         annotations:
 197           temperature_celsius: "{{ with printf \"ipmi_temperature_celsius{instance='%s',id='%s'}\" $labels.instance $labels.id | query }}{{ . | first | value | humanize }}C{{end}}"
 198       - alert: ipmi voltage alarm
 199         expr: ipmi_voltage_state > 0
 200         for: 5m
 201         labels:
 202           alertgroup: "{{ $labels.instance }}"
 203         annotations:
 204           voltage_volts: "{{ with printf \"ipmi_voltage_volts{instance='%s',id='%s'}\" $labels.instance $labels.id | query }}{{ . | first | value | humanize }}V{{end}}"
 205       - alert: ipmi power alarm
 206         expr: ipmi_power_state > 0 or ipmi_sensor_state{type=~"Power .*"} > 0
 207         for: 5m
 208         labels:
 209           alertgroup: "{{ $labels.instance }}"
 210   - name: juniper
 211     rules:
 212       - alert: juniper fan alarm
 213         expr: jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!="running"} > 0
 214         for: 5m
 215         labels:
 216           alertgroup: "{{ $labels.site }}"
 217       - alert: juniper power alarm
 218         expr: jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!="running"} > 0
 219         for: 5m
 220         labels:
 221           alertgroup: "{{ $labels.site }}"
 222   - name: mail
 223     rules:
 224       - alert: exim queue length
 225         expr: exim_queue > exim_queue_limit
 226         for: 60m
 227         labels:
 228           alertgroup: mail
 229         annotations:
 230           queue_length: "{{ $value }}"
 231       - alert: mailman queue length
 232         expr: mailman_queue_length > 200
 233         for: 60m
 234         labels:
 235           alertgroup: mail
 236         annotations:
 237           queue_length: "{{ $value }}"
 238   - name: mdadm
 239     rules:
 240       - alert: mdadm array inactive
 241         expr: node_md_state{state="inactive"} > 0
 242         for: 0m
 243         labels:
 244           alertgroup: "{{ $labels.instance }}"
 245         annotations:
 246           required: "{{ with printf \"node_md_disks_required{instance='%s',device='%s'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 247           active: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='active'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 248           failed: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='failed'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 249           spare: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='spare'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 250       - alert: mdadm array degraded
 251         expr: sum (node_md_disks{state="active"}) without (state) < node_md_disks_required
 252         for: 0m
 253         labels:
 254           alertgroup: "{{ $labels.instance }}"
 255         annotations:
 256           required: "{{ with printf \"node_md_disks_required{instance='%s',device='%s'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 257           active: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='active'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 258           failed: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='failed'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 259           spare: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='spare'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 260       - alert: mdadm disk failed
 261         expr: node_md_disks{state="failed"} > 0
 262         for: 0m
 263         labels:
 264           alertgroup: "{{ $labels.instance }}"
 265         annotations:
 266           required: "{{ with printf \"node_md_disks_required{instance='%s',device='%s'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 267           active: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='active'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 268           failed: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='failed'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 269           spare: "{{ with printf \"node_md_disks{instance='%s',device='%s',state='spare'}\" $labels.instance $labels.device | query }}{{ . | first | value | humanize }} disks{{end}}"
 270   - name: memory
 271     rules:
 272       - alert: low memory
 273         expr: node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes < 0.1
 274         for: 15m
 275         labels:
 276           alertgroup: "{{ $labels.instance }}"
 277         annotations:
 278           memory_free: "{{ $value | humanizePercentage }}"
 279       - alert: memory pressure
 280         expr: rate(node_pressure_memory_waiting_seconds_total[5m]) > 0.6
 281         for: 60m
 282         labels:
 283           alertgroup: "{{ $labels.instance }}"
 284         annotations:
 285           pressure: "{{ $value | humanizePercentage }}"
 286       - alert: oom kill detected
 287         expr: increase(node_vmstat_oom_kill[1m]) > 0
 288         for: 0m
 289         labels:
 290           alertgroup: "{{ $labels.instance }}"
 291         annotations:
 292           new_oom_kills: "{{ $value }}"
 293   - name: network
 294     rules:
 295       - alert: interface transmit rate
 296         expr: rate(node_network_transmit_bytes_total[1m]) / node_network_speed_bytes > 0.98
 297         for: 5m
 298         labels:
 299           alertgroup: "{{ $labels.instance }}"
 300         annotations:
 301           bandwidth_used: "{{ $value | humanizePercentage }}"
 302       - alert: interface receive rate
 303         expr: rate(node_network_receive_bytes_total[1m]) / node_network_speed_bytes > 0.98
 304         for: 5m
 305         labels:
 306           alertgroup: "{{ $labels.instance }}"
 307         annotations:
 308           bandwidth_used: "{{ $value | humanizePercentage }}"
 309       - alert: interface transmit errors
 310         expr: rate(node_network_transmit_errs_total[1m]) / rate(node_network_transmit_packets_total[1m]) > 0.01
 311         for: 5m
 312         labels:
 313           alertgroup: "{{ $labels.instance }}"
 314         annotations:
 315           error_rate: "{{ $value | humanizePercentage }}"
 316       - alert: interface receive errors
 317         expr: rate(node_network_receive_errs_total[1m]) / rate(node_network_receive_packets_total[1m]) > 0.01
 318         for: 5m
 319         labels:
 320           alertgroup: "{{ $labels.instance }}"
 321         annotations:
 322           error_rate: "{{ $value | humanizePercentage }}"
 323       - alert: conntrack entries
 324         expr: node_nf_conntrack_entries / node_nf_conntrack_entries_limit > 0.8
 325         for: 5m
 326         labels:
 327           alertgroup: "{{ $labels.instance }}"
 328         annotations:
 329           entries_used: "{{ $value | humanizePercentage }}"
 330   - name: planet
 331     rules:
 332       - alert: planet dump overdue
 333         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/(pbf|planet)/.*"} > 7 * 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 334         for: 24h
 335         labels:
 336           alertgroup: planet
 337         annotations:
 338           overdue_by: "{{ $value | humanizeDuration }}"
 339       - alert: notes dump overdue
 340         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/notes/.*"} > 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 341         for: 6h
 342         labels:
 343           alertgroup: planet
 344         annotations:
 345           overdue_by: "{{ $value | humanizeDuration }}"
 346       - alert: daily replication feed delayed
 347         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/day/.*"} > 86400 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 348         for: 3h
 349         labels:
 350           alertgroup: planet
 351         annotations:
 352           delayed_by: "{{ $value | humanizeDuration }}"
 353       - alert: hourly replication feed delayed
 354         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/hour/.*"} > 3600 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 355         for: 30m
 356         labels:
 357           alertgroup: planet
 358         annotations:
 359           delayed_by: "{{ $value | humanizeDuration }}"
 360       - alert: minutely replication feed delayed
 361         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/minute/.*"} > 60 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 362         for: 5m
 363         labels:
 364           alertgroup: planet
 365         annotations:
 366           delayed_by: "{{ $value | humanizeDuration }}"
 367       - alert: changeset replication feed delayed
 368         expr: time() - file_stat_modif_time_seconds{path=~"/store/planet/replication/changesets/.*"} > 60 and ignoring (job, name, path) chef_role{name="planetdump"} == 1
 369         for: 5m
 370         labels:
 371           alertgroup: planet
 372         annotations:
 373           delayed_by: "{{ $value | humanizeDuration }}"
 374   - name: postgresql
 375     rules:
 376       - alert: postgresql down
 377         expr: pg_up == 0
 378         for: 1m
 379         labels:
 380           alertgroup: "{{ $labels.instance }}"
 381       - alert: postgresql replication delay
 382         expr: pg_replication_lag_seconds > 5
 383         for: 1m
 384         labels:
 385           alertgroup: "{{ $labels.instance }}"
 386         annotations:
 387           delay: "{{ $value | humanizeDuration }}"
 388       - alert: postgresql connection limit
 389         expr: sum (pg_stat_activity_count) by (instance, server) / sum (pg_settings_max_connections) by (instance, server) > 0.8
 390         for: 1m
 391         labels:
 392           alertgroup: "{{ $labels.instance }}"
 393         annotations:
 394           connections_used: "{{ $value | humanizePercentage }}"
 395       - alert: postgresql deadlocks
 396         expr: increase(pg_stat_database_deadlocks[1m]) > 5
 397         for: 0m
 398         labels:
 399           alertgroup: "{{ $labels.instance }}"
 400         annotations:
 401           new_deadlocks: "{{ $value }}"
 402       - alert: postgresql slow queries
 403         expr: pg_slow_queries > 0
 404         for: 5m
 405         labels:
 406           alertgroup: "{{ $labels.instance }}"
 407         annotations:
 408           queries: "{{ $value }}"
 409   - name: smart
 410     rules:
 411       - alert: smart failure
 412         expr: smart_health_status == 0
 413         for: 60m
 414         labels:
 415           alertgroup: "{{ $labels.instance }}"
 416       - alert: smart ssd wearout approaching
 417         expr: smart_percentage_used >= 90
 418         for: 60m
 419         labels:
 420           alertgroup: "{{ $labels.instance }}"
 421         annotations:
 422           percentage_used: "{{ $value | humanizePercentage }}"
 423   - name: ssl
 424     rules:
 425       - alert: ssl certificate probe failed
 426         expr: ssl_probe_success == 0
 427         for: 60m
 428         labels:
 429           alertgroup: ssl
 430       - alert: ssl certificate expiry
 431         expr: ssl_verified_cert_not_after{chain_no="0"} - time() < 86400 * 14
 432         for: 0m
 433         labels:
 434           alertgroup: ssl
 435         annotations:
 436           expires_in: "{{ $value | humanizeDuration }}"
 437       - alert: ssl certificate revoked
 438         expr: ssl_ocsp_response_status == 1
 439         for: 0m
 440         labels:
 441           alertgroup: ssl
 442       - alert: ocsp status unknown
 443         expr: ssl_ocsp_response_status == 1
 444         for: 0m
 445         labels:
 446           alertgroup: ssl
 447   - name: statuscake
 448     rules:
 449       - alert: statuscake uptime check failing
 450         expr: statuscake_uptime{status="down",paused="false"} > 0
 451         for: 0m
 452         labels:
 453           alertgroup: statuscake
 454   - name: systemd
 455     rules:
 456       - alert: systemd failed service
 457         expr: node_systemd_unit_state{state="failed",name!="chef-client.service"} == 1
 458         for: 5m
 459         labels:
 460           alertgroup: "{{ $labels.instance }}"
 461       - alert: systemd failed service
 462         expr: node_systemd_unit_state{state="failed",name="chef-client.service"} == 1
 463         for: 6h
 464         labels:
 465           alertgroup: "{{ $labels.instance }}"
 466   - name: tile
 467     rules:
 468       - alert: renderd replication delay
 469         expr: renderd_replication_delay > 120
 470         for: 15m
 471         labels:
 472           alertgroup: tile
 473         annotations:
 474           delay: "{{ $value | humanizeDuration }}"
 475       - alert: missed tile rate
 476         expr: sum(rate(modtile_http_response_total{code="404"}[5m])) by (instance) / sum(rate(modtile_http_response_total[5m])) by (instance) > 0.05
 477         for: 5m
 478         labels:
 479           alertgroup: tile
 480         annotations:
 481           miss_rate: "{{ $value | humanizePercentage }}"
 482   - name: time
 483     rules:
 484       - alert: clock not synchronising
 485         expr: min_over_time(node_timex_sync_status[1m]) == 0 and node_timex_maxerror_seconds >= 16
 486         for: 5m
 487         labels:
 488           alertgroup: "{{ $labels.instance }}"
 489       - alert: clock skew detected
 490         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)
 491         for: 5m
 492         labels:
 493           alertgroup: "{{ $labels.instance }}"
 494         annotations:
 495           skew: "{{ with printf \"node_timex_offset_seconds{instance='%s'}\" $labels.instance | query }} {{ . | humanizeDuration }}{{ end }}"
 496   - name: web
 497     rules:
 498       - alert: web error rate
 499         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
 500         for: 5m
 501         labels:
 502           alertgroup: web
 503         annotations:
 504           error_rate: "{{ $value | humanizePercentage }}"
 505       - alert: job processing rate
 506         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
 507         for: 15m
 508         labels:
 509           alertgroup: web
 510         annotations:
 511           job_processing_rate: "{{ $value | humanizePercentage }}"