pgwatch2 1.12.0: periodically msg "WARN: main: Could not find PG version info for DB XXX, skipping ..." when monitoring aurora databases

[jjamor]

Hello,

We have pgwatch2 1.12..0 monitoring several AWS (aurora), Azure and on-prem postgres databases.

Periodically, some aurora DBs stop being monitored with the given log in pgwatch2:

"WARN: main: Could not find PG version info for DB XXX, skipping ..."

This is recovered when restarting pgwatch2.

We are upgrading from pgwatch2 1.8.5, and this issue occurs only with the latest 1.12.0, with the same configuration.

Is this a known error in 1.12.0?

Many thanks in advance.

[pashagolub]

Hi,

Seems like issued by this line. One of the reason is changes in metrics and/or monitored databases.

This is not known issue. Maybe you can provide some context?

[jjamor]

It happens randomly with AWS aurora databases.

When the error occurs,, if I try to manually open a connection from the same machine (using psql client, for example), the connection can be establihshed without problems. Also, the error is fixed after restarting pgwatch2.

Is there a way to force a reconnection when the current connection is lost, without restarting pgwatch2?

Thanks!

[pashagolub]

What are the database name(s) of affected sources?
There is no connection loss. There is an absence of version (meta) information about the monitored databases.

[pashagolub]

And what are select version() results?

[jjamor]

Hello, sorry for the delay,

The problem happens randomly with AWS Aurora databases, it is fixed only by restarting pgwatch2 daemon. When a monitored database does not respond to pgwatch2, I can however connect to it with psql and show version:

`select version();
version

PostgreSQL 14.6 on aarch64-unknown-linux-gnu, compiled by aarch64-unknown-linux-gnu-gcc (GCC) 7.4.0, 64-bit`