Version Information

Server Version:
v2.13.0-cloud.1

Environment

Cloud

What is the current behaviour?

I have two Hasura instances which differ in how they have constructed the hdb_catalog in the database.

If I run this sql in both:

SELECT * FROM information_schema.tables 
WHERE table_schema = 'hdb_catalog'

I see different results. The first is what we have (rightly or wrongly) considered to be the 'correct' behaviour (the behaviour we have coded against)

The second instance does not surface these tables, instead (for the same query) giving us:

I'm not sure which is the intended behaviour. The reason this is an issue is because we have a script which can return the number of unprocessed events at any one time:

post(HASURA_QUERY_ENDPOINT, {
      headers: {"X-Hasura-Admin-Secret": HASURA_ADMIN_SECRET},
      body: JSON.stringify({
         type: "run_sql",
         args: { sql: "select count(*) from hdb_catalog.hdb_scheduled_events where status = 'scheduled'" },
      }),
});

This obviously throws an error in the second example as the table does not exist. I would like to understand:

1. Which of these two results is intended/expected?
2. How/where are events stored in the second example?
3. How can I go about making these consistent?

Thanks in advance

(Related: https://stackoverflow.com/q/70010557/3001761)