PG_STAT_REPLICATION

startup: The thread is being started.
catchup: The thread is establishing a connection between the standby node and the primary node.
streaming: The thread has established a connection between the standby node and the primary node and is replicating data streams.
backup: The thread is sending a backup.
stopping: The thread is being stopped.

PG_STAT_REPLICATION displays information about the log synchronization thread, such as the locations where the sender sends logs and where the receiver receives logs.

**Table 1** PG_STAT_REPLICATION columns
Name	Type	Description
pid	bigint	PID of the thread.
usesysid	oid	User system ID.
usename	name	Username.
application_name	text	Program name.
client_addr	inet	Client address.
client_hostname	text	Client name.
client_port	integer	Port of the client.
backend_start	timestamp with time zone	Start time of the program.
state	text	Status of the log synchronization thread: startup: The thread is being started. catchup: The thread is establishing a connection between the standby node and the primary node. streaming: The thread has established a connection between the standby node and the primary node and is replicating data streams. backup: The thread is sending a backup. stopping: The thread is being stopped.
sender_sent_location	text	Location where the sender sends logs.
receiver_write_location	text	Location where the receiver writes logs.
receiver_flush_location	text	Location where the receive end flushes logs.
receiver_replay_location	text	Location where the receive end replays logs.
sync_priority	integer	Priority of synchronous duplication (0 indicates asynchronization).
sync_state	text	Synchronization state: async: asynchronous replication. sync: synchronous replication. potential: The standby node is asynchronous currently, but if a current synchronization node fails, the standby node becomes synchronous. qurom: switches between synchronous and asynchronous state to ensure that there are more than a certain number of synchronous standby nodes. Generally, the number of synchronous standby nodes is (n+1)/2-1, where n indicates the total number of copies. Whether the standby node is synchronous depends on whether logs are received first. For details, see the description of the synchronous_standby_names parameter.