plc-mirror/metrics/prometheus/exporters/query-exporter/config.yaml

databases:
  db1:
    dsn: env:POSTGRES_URL

metrics:
  repos_fully_indexed:
    type: gauge
    description: Repositories fully indexed
  repos_seen:
    type: gauge
    description: Repositories seen
  repos_failed:
    type: gauge
    description: Repositories that we failed to index
  consumer_bad_records:
    type: gauge
    description: Records received from firehose that we failed to process
    labels: [pds, error]
  # posts_lang:
  #   type: summary
  #   description: Posts by language
  #   labels: [uk, lt, et, lv, pl, ga, fi, sv,
  #   en, jp, de, fr, pt, es, nl, ko, tr, zh, ru]

queries:
  query1:
    interval: 30
    databases: [db1]
    metrics: [repos_fully_indexed]
    sql: >
      select count(*) as repos_fully_indexed
        from repos left join pds on repos.pds = pds.id
        where failed_attempts < 3
          and last_indexed_rev <> ''
          and (last_indexed_rev >= first_rev_since_reset
                or first_rev_since_reset is null or first_rev_since_reset = '')
          and (repos.first_cursor_since_reset >= pds.first_cursor_since_reset
                or repos.first_cursor_since_reset is null or repos.first_cursor_since_reset = 0);
  query2:
    interval: 30
    databases: [db1]
    metrics: [repos_seen]
    sql: select count(*) as repos_seen from repos;
  query3:
    interval: 30
    databases: [db1]
    metrics: [repos_failed]
    sql: select count(*) as repos_failed from repos where failed_attempts >= 3;
  # query4:
  #   interval: 300
  #   databases: [db1]
  #   metrics: [posts_lang]
  #   sql: select count(*) as uk from records where collection in ('app.bsky.feed.post') and content::text like '%"langs": ["uk"]%';
  bad_records:
    interval: 30
    databases: [db1]
    metrics: [consumer_bad_records]
    sql: |
      select count(*) as consumer_bad_records, host as pds, error
        from (
          select id, created_at, pds, cursor, content,
              regexp_replace(regexp_replace(regexp_replace(error,
                'did:[\:a-z0-9]+', 'did:xxx', 'g'),
                'json\.RawMessage\{[^}]+\}', 'json.RawMessage{...}', 'g'),
                '[0-9]{1,3}(\.[0-9]{1,3}){3}\:[0-9]+', '<IP>\:<port>', 'g') as error
          from bad_records
          )
        join
          pds
        on pds=pds.id
        group by error, host;