Do I need to restart the postgresql service for log_checkpoints to make it active . I would like to enable this parameter temporarily and check it?
Nope, just get it in postgresql.conf and reload it. i.e. pg_ctl reload
postgres=# show max_wal_size;
max_wal_size
--------------
2GB
(1 row)
With the deletes, vacuum full, and reindex it is generating WALs at a rapid rate thus trigging another checkpoint given the 2GB size. The vacuum full and reindex are most likely causing Zabbix to be blocked.
I would rethink the vacuum full and reindex; before changing max_wal_size.