Postgres - APG - Trigger chaining is slow

I have two tables, T1 and T2 with 1 trigger on each. Following occurs: -- Execute Query 1, say Q1

UPDATE Y.T1
    SET name='name'
  WHERE id=0 ;

this results in trigger (Trigger1) firing query:

UPDATE Y.T2   SET IsFlat = 0 where id = -1;

---- few seconds later, i execute query Q2:

WITH X AS ( 
 select TV0.Id, row_number() over(  
			 partition by coalesce(TV0.ParentId, 0)
			 order by UPPER(NAME)  
				) as RN
from Y.T2 TV0
where TV0.TagId = 604 
)  
update Y.T2 TV
set SortOrder = X.RN
from X
where X.Id = TV.Id ; 

this results in trigger (Trigger2) being fired:

begin
        -- get timestamp say Time1
       for rec in SELECT n.id, n.tagId, n.parentId as newParentId, o.parentId as oldParentId
            FROM new_table n JOIN old_table o on n.id = o.id
            and o.parentId != n.parentId                   
        loop. -- it never enters this loop, mycounter when printed always returns 0
      
            mycounter:=mycounter+1;
        end loop;
     -- get Timestamp say Time2
end;

Query 2 takes 30 seconds If Q1 precedes it. Otherwise it finishes in less than 2 sec. Query 2 takes 30 seconds because of the trigger Trigger2, which consumes almost 28-30 sec to execute. Query 2 by itself finishes in 2 sec when not preceded by Query1. Execution plan shows trigger taking time ~28 sec. If i replace the join in trigger

FROM new_table n JOIN old_table o

to use either of one table, which eliminates join, query comes back in ~2-3 sec. So we know join (defined in trigger) is slow if it succeeds Query 1. Why would that be the case?

I also tried update statement to replace Query 2 to:

update Y.T2 TV
set SortOrder = 1
where SortOrder = 1;

and query join finishes in time. I add another trigger with same join condition and query time increases to 2X say ~60 sec in this case.