David Steele

OK, now I see that you have set `archive-missing-retry=n`. Since `process-max=4` WAL segments are being fetched out of order and sometimes an earlier segment is the last to arrive. Since...

Hmmm, no, ignore the above. Even when `archive-missing-retry=n` we'll only return when the WAL file is verified as missing, i.e. there is an `ok` file but no WAL segment so...

> I'll dig into how the ok files are generated. I can't see anything problematic in how these are generated. Can we get the archive-async logs? Hopefully they are at...

These are the archive logs. We are looking for the archive async logs, which would be in `prod-pg-f3c082-archive-get-async.log`. They should contain this string: `archive-get:async command begin`.

Closing this since there has been no response with logs.

> The error from S3 (AWS in this case) was not consumed appropriately by pgbackrest and instead of failing Actually, we do fail here but Postgres treats any failure as...

I'm still not convinced that pgAudit needs to compensate for odd things that other extensions do. Seems like a rabbit hole we don't want to go down.

This will be addressed by PR #272.

We do updates for the new version of Postgres when the first beta comes out in May. We do not try to track development as it happens.

> What pgaudit/PostgreSQL branches should them use? You can build the pgAudit main (or REL_17_STABLE) branch against the Postgres REL_17_STABLE branch.