Module catalog_serving

Special cases related to the “catalog serving” of Materialize

Every Materialize deployment has a pre-installed mz_catalog_server cluster, which has several indexes to speed up common catalog queries. We also have a special mz_support role, which can be used by support teams to diagnose a deployment. For each of these use cases, we have some special restrictions we want to apply. The logic around these restrictions is defined here.

Functions

auto_run_on_catalog_server

Checks whether or not we should automatically run a query on the mz_catalog_server cluster, as opposed to whatever the current default cluster is.

check_cluster_restrictions

Checks if we’re currently running on the MZ_CATALOG_SERVER_CLUSTER, and if so, do we depend on any objects that we’re not allowed to query from the cluster.