pg_locks

Hi,
i would prefer to reduce the amount of data and skip lock modes that are absent but add granted and locktype to pg_locks:

pg_lock:
  name: pg_lock
  desc: PostgreSQL lock distribution by mode and database incl. granted
  query: |-
    SELECT d.datname, locktype, mode, granted, count(*) AS count
        FROM pg_locks JOIN pg_database d on d.oid=database
        WHERE DATABASE IS NOT NULL
        GROUP BY 1, 2, 3, 4 ;

  min_version: 090400
  tags: [ cluster ]
  metrics:
    - datname: { usage: LABEL ,description: "Name of the database this backend is connected to" }
    - locktype: { usage: LABEL ,description: "Locktype" }
    - mode:    { usage: LABEL ,description: "Name of the lock mode held or desired by this process" }
    - granted: { usage: LABEL ,description: "granted" }
    - count:   { usage: GAUGE ,description: "Number of locks of corresponding mode and database" }

Locate the existing pg_lock metric definition and review how its current pg_locks query is configured. Update it to group by locktype, mode, and granted while omitting absent lock modes and excess data; done means the exported metric has the requested labels and count for each database.