Cloud

System Virtual Tables

System virtual tables in the system schema expose cluster runtime information and external-source metadata. They provide SQL-queryable alternatives to SHOW statements, with support for WHERE, ORDER BY, and LIMIT clauses.

System virtual tables filter rows by the caller’s privileges. Cluster runtime tables (system.nodes, system.queries, system.execs) are visible only to a superuser. Tables that expose catalog and connection metadata filter rows by the caller’s grants on the database, namespace (also called schema), and metadata.

Available tables

Table Equivalent statement Description

system.catalogs

SHOW CATALOGS

All catalogs (both Redpanda and Iceberg) visible to the caller.

system.iceberg_catalogs

SHOW ICEBERG CATALOGS

Iceberg catalogs with their REST endpoint, warehouse, and authentication type.

system.iceberg_tables

SHOW ICEBERG TABLES

Iceberg tables whose schema has been loaded into the local catalog.

system.kafka_connections

SHOW REDPANDA/KAFKA CATALOGS

Redpanda catalog connections, including any linked Iceberg catalog.

system.kafka_sources

SHOW REDPANDA/KAFKA TABLES

Tables mapped to Redpanda topics through a Redpanda catalog.

system.nodes

SHOW NODES

Current state of all nodes in the cluster.

system.queries

SHOW QUERIES

Currently running queries.

system.execs

SHOW EXECS

Currently running execution tasks.

system.catalogs

All catalogs in the cluster (both Redpanda and Iceberg), across databases and schemas visible to the caller.

Column Type Nullable Description

name

text

No

Catalog name.

namespace_name

text

No

Schema containing the catalog.

type

text

No

Catalog type: REDPANDA or ICEBERG.

 
SELECT * FROM system.catalogs;

For per-catalog detail, query system.iceberg_catalogs or system.kafka_connections.

system.iceberg_catalogs

Iceberg catalog connections, with their REST endpoint and authentication details.

Column Type Nullable Description

name

text

No

Iceberg catalog name.

uri

text

No

REST catalog endpoint URI.

warehouse

text

No

Iceberg warehouse identifier or location.

auth_type

text

No

Authentication type for the REST catalog. One of oauth2, basic, aws_sigv4, or empty if the catalog connects without authentication.

namespace_name

text

No

Schema containing the catalog.

database_name

text

No

Database containing the catalog.

 
SELECT name, uri, auth_type FROM system.iceberg_catalogs;

system.iceberg_tables

Iceberg tables whose schema has been loaded into the local catalog. Schemas are loaded by an explicit REFRESH statement or auto-loaded when you CREATE TABLE against an Iceberg catalog. See Query Iceberg-enabled Topics for details on when to refresh. Each row represents one root Iceberg table.

Column Type Nullable Description

database_name

text

No

Database containing the registered table.

namespace_name

text

No

Schema containing the registered table.

catalog_name

text

No

Iceberg catalog the table belongs to.

name

text

No

Qualified Iceberg table path (for example, ns1.ns2.tbl).

oid

int

No

Object identifier of the table type. Joinable with pg_type.oid.

Stale entries: dropping an Iceberg table from the REST catalog leaves its registration in the Redpanda SQL catalog until you drop it explicitly. The view continues to list it. This mirrors system.kafka_sources behavior.

 
SELECT * FROM system.iceberg_tables WHERE catalog_name = 'lakehouse_catalog';

system.kafka_connections

Redpanda catalog connections, including any linked Iceberg catalog for queries that span live records and Iceberg-translated history. For details on linking and Iceberg-translated history, see ALTER REDPANDA CATALOG.

Column Type Nullable Description

database_name

text

No

Database containing the connection.

namespace_name

text

No

Schema containing the connection.

name

text

No

Catalog (connection) name.

options

text

No

Connection options as a comma-separated string of key: value pairs. Includes broker addresses, Kafka timeout, Schema Registry endpoint, Schema Registry timeout, and a nested librdkafka configuration list in [key=value, …​] form. Pandaproxy options are included when configured.

iceberg_catalog

text

Yes

Linked Iceberg catalog as schema.name, or NULL if no link is configured. Returns NULL when the caller doesn’t have CONNECT on the linked catalog’s database and USAGE on its schema (or an ANY grant on the metadata).

 
SELECT name, iceberg_catalog FROM system.kafka_connections;

system.kafka_sources

Tables mapped to Redpanda topics through a Redpanda catalog. One row per CREATE TABLE catalog⇒name WITH (…​) mapping.

Column Type Nullable Description

database_name

text

No

Database containing the source table.

namespace_name

text

No

Schema containing the source table.

name

text

No

Table name.

connection_name

text

No

Redpanda catalog (connection) the table belongs to.

topic_name

text

No

Source topic name.

subject_name

text

No

Schema Registry subject name. Empty if not configured.

lookup_policy

text

No

Schema lookup policy. See CREATE TABLE.

error_handling_policy

text

No

Deserialization error-handling policy. See CREATE TABLE.

struct_mapping_policy

text

No

Nested-struct mapping policy. See CREATE TABLE.

output_schema_full_message_name

text

No

Full Protobuf message name. Empty if not configured.

 
SELECT name, topic_name, subject_name FROM system.kafka_sources
WHERE connection_name = 'production_redpanda';

system.nodes

Current state of each node in the cluster. Visible only to a superuser. For the column reference, see SHOW NODES.

SELECT * FROM system.nodes;

system.queries

Currently running queries on the cluster. Visible only to a superuser.

Column Type Nullable Description

qid

text

No

Query identifier.

requester

text

No

Identifier of the node that received the query from the client.

scheduler

text

No

Identifier of the node scheduling the query’s execution tasks.

workers

bigint

No

Number of worker tasks assigned to the query.

state

text

No

Current state of the query.

created

timestamp

No

Time the query was received.

accepted

timestamp

Yes

Time the query was accepted for execution. NULL until accepted.

scheduled

timestamp

Yes

Time the query’s execution tasks were scheduled. NULL until scheduled.

executed

timestamp

Yes

Time execution started. NULL until execution starts.

finished

timestamp

Yes

Time execution finished. NULL while the query is still running.

 
SELECT qid, state, created FROM system.queries WHERE state <> 'FINISHED';

system.execs

Currently running execution tasks across the cluster. Visible only to a superuser. Each query in system.queries typically corresponds to one or more rows in this table.

Column Type Nullable Description

node

text

No

Node executing the task.

qid

text

No

Identifier of the parent query. Join with system.queries.qid.

data_task_id

bigint

Yes

Identifier of the data task within the query.

state

text

No

Current state of the execution task.

memory

bigint

No

Memory consumed by the task, in bytes.

privileged

boolean

No

TRUE if the task runs with privileged access.

 
SELECT node, qid, state, memory FROM system.execs ORDER BY memory DESC;