Verified Commit 56a7fe90 authored by anarcat's avatar anarcat 💥
Browse files

extend udldap refresh checks 

When I have a hour-long downtime at home (which happens), this trips
needlessly.

This is just a sanity check, it's fine to wait a day.
parent 524d63a4

rules.d/tpa_node.rules

+3 −4
Original line number Diff line number Diff line
@@ -321,13 +321,12 @@ groups: 
      playbook: "TODO"



  - alert: UDLdapLastRefreshTooOld

    expr: (time() - node_path_modification_timestamp_seconds) > 60*60

    expr: (time() - node_path_modification_timestamp_seconds) > 24*60*60

    labels:

      severity: warning

    annotations:

      summary: "userdir-ldap has not refreshed correctly on {{ $labels.alias }} for more than 1h"

      summary: "userdir-ldap has not refreshed correctly on {{ $labels.alias }} for too long"

      description: |

        On {{ $labels.alias }} the file {{ $labels.path }} has

        a last modification date that's one hour or more in the past. The userdir-ldap

        replication should be verified.

        a last modification date that is {{ $value | humanizeDuration }} old. The userdir-ldap replication should be verified.

      playbook: "TODO"

tests/tpa_node.yml

+3 −4
Original line number Diff line number Diff line
@@ -659,7 +659,7 @@ tests: 
    input_series:

      # file too old, firing

      - series: 'node_path_modification_timestamp_seconds{alias="submit-01.torproject.org",instance="submit-01.torproject.org:9100",job="node",path="/var/lib/misc/thishost/last_update.trace",team="TPA"}'

        values: '-3601'

        values: '-86401'

      # file is recent enough, not firing

      - series: 'node_path_modification_timestamp_seconds{alias="web-fsn-02.torproject.org",instance="web-fsn-02.torproject.org:9100",job="node",path="/var/lib/misc/thishost/last_update.trace",team="TPA"}'

        values: '-150'
@@ -675,9 +675,8 @@ tests: 
              job: "node"

              path: "/var/lib/misc/thishost/last_update.trace"

            exp_annotations:

              summary: "userdir-ldap has not refreshed correctly on submit-01.torproject.org for more than 1h"

              summary: "userdir-ldap has not refreshed correctly on submit-01.torproject.org for too long"

              description: |

                On submit-01.torproject.org the file /var/lib/misc/thishost/last_update.trace has

                a last modification date that's one hour or more in the past. The userdir-ldap

                replication should be verified.

                a last modification date that is 1d 0h 0m 1s old. The userdir-ldap replication should be verified.

              playbook: "TODO"