Hi, On Tue, Jan 20, 2026 at 7:21 AM Michael Paquier <[email protected]> wrote: > > On Mon, Jan 19, 2026 at 11:35:28PM +0800, Xuneng Zhou wrote: > > On Mon, Jan 19, 2026 at 8:13 AM Michael Paquier <[email protected]> wrote: > >> I think that > >> there would be a point in expanding the SQL functions to report more > >> states of the startup process, including the data received by the > >> startup process, but we should not link that to the state of the WAL > >> receiver. An extra reason to not do that: WAL receivers are not the > >> only source feeding data to the startup process, we could have data > >> pushed to pg_wal/, or archive commands/modules doing this job. > > > > +1. I'll prepare a separate patch to expose startup process metrics > > like pg_stat_get_wal_receiver does. This would complement > > pg_stat_wal_receiver without coupling the two subsystems. > > In this area, I mean to expose the contents of XLogRecoveryCtlData at > SQL level. It may be better to move this structure to a header, and > have the new SQL function in xlogfuncs.c. That's at least how I would > shape such a change. > -- > Michael
Thanks for the suggestion! It makes sense to me. -- Best, Xuneng
