Exporter Errors for Two Metrics

[kyle-rf]

What did you do?
Upgraded postgres operator to 4.7.4
Upgraded a running cluster

What did you expect to see?
Continued functioning as before upgrade

What did you see instead? Under which circumstances?
The exporter container is logging the following errors:
"Error running query on database "127.0.0.1:5432": ccp_pg_stat_statements_reset pq: function monitor.pg_stat_statements_reset_info(integer) does not exist"

"Error running query on database "127.0.0.1:5432": ccp_nodemx_data_disk pq: query-specified return tuple and function return type are not compatible"

Environment

• System information:

  Linux 5.4.0-70-generic x86_64

• postgres_exporter version:

	registry.developers.crunchydata.com/crunchydata/crunchy-postgres-exporter:centos8-4.7.4

• postgres_exporter flags:

I don't what exporter flags are

• PostgreSQL version:

        13.5
	registry.developers.crunchydata.com/crunchydata/crunchy-postgres-ha:centos8-13.5-4.7.4

• Logs:

"Error running query on database "127.0.0.1:5432": ccp_pg_stat_statements_reset pq: function monitor.pg_stat_statements_reset_info(integer) does not exist"

"Error running query on database "127.0.0.1:5432": ccp_nodemx_data_disk pq: query-specified return tuple and function return type are not compatible"

[keithf4]

You may need to open an issue with the Operator team. https://github.com/CrunchyData/postgres-operator

That monitor.pg_stat_statements_reset_info function is handled by the setup file in pgmonitor

https://github.com/CrunchyData/pgmonitor/blob/main/postgres_exporter/common/pg13/setup.sql#L423

And the pg_nodemx functions are specific to the container environment and are part of this extension - https://github.com/CrunchyData/pgnodemx

How the operator takes care of updating these objects and extensions as part of its updates are specific to the operator itself.