Horton

Re: [Proposal] Expose internal MultiXact member count function for efficient monitoring

Michael Paquier <michael@paquier.xyz>

View thread
From: Michael Paquier <michael@paquier.xyz>
To: Naga Appani <nagnrik@gmail.com>
Cc: Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>, Tomas Vondra <tomas@vondra.me>, Xuneng Zhou <xunengzhou@gmail.com>, torikoshia <torikoshia@oss.nttdata.com>, Kirill Reshke <reshkekirill@gmail.com>, pgsql-hackers@postgresql.org
Date: 2025-12-17T04:19:11Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Add pg_get_multixact_stats()

  2. Add MultiXactOffsetStorageSize() to multixact_internal.h

  3. Change GetMultiXactInfo() to return the next multixact offset

  4. Widen MultiXactOffset to 64 bits

  5. Refactor ReadMultiXactCounts() into GetMultiXactInfo()

  6. Move SQL-callable code related to multixacts into its own file

  7. Split func.sgml into more manageable pieces

On Sat, Dec 13, 2025 at 01:34:47PM -0600, Naga Appani wrote:
> Documentation changes:
> - Removed the NULL-return discussion from func-info.sgml, as the
>   statistics are now always available.
> - Updated maintenance.sgml to clarify that exceeding the historical
>   2^32 member limit no longer causes wraparound, but instead triggers
>   more aggressive vacuum activity for disk space management.
> 
> I validated the behavior before and after cleanup.
> The function correctly reports current usage (beyond the old limits) and
> resets once multixacts are removed:

+	/*
+	 * Calculate storage space for members. Members are stored in groups,
+	 * with each group containing MULTIXACT_MEMBERS_PER_MEMBERGROUP members
+	 * and taking MULTIXACT_MEMBERGROUP_SIZE bytes.
+	 */
+	membersBytes = (int64) (members / MULTIXACT_MEMBERS_PER_MEMBERGROUP) *
+				   MULTIXACT_MEMBERGROUP_SIZE;

This is the key point of the patch internal logic.  And there is one
thing that I am wondering here.  The amount of space taken by a number
of members depends on the other compiled constants from
multixact_internal.h.  Hence, rather than calculate the amount of
space taken by a set of members in some code hidden in the SQL
function, could it be better to put that directly as a macro or an
inline function in multixact_internal.h?
--
Michael