Too many glob / SQL ?

Is there a limit on too much globbing? and SQL?

I've basically got a LONG process that repeats monthly. Every month, task 1
is created.

task 1 >> creates 20 task 2 >> each creates 1-20 task 3 etc >> each creates
task 4 etc ..

If there are no task 4s, it checks task 3, etc - going up until all tasks
are completed...

every 5 minutes, 1 task is completed.

ISSUE:  One of these jobs is a 'jobs' task these are stored in the jobs
directory    path/.../data/*/*/Jobs/(jobs info here).txt

PART 1 QUESTION:

Instead of checking the 'jobs' directory one at a time (5 minutes each),
I'm thinking of doing a GLOB - to locate the 'jobs' directory and create a
task ONLY if the directory exists. instead of creating a jobs check each
and every time.

If I create a GLOB to locate the jobs directory and create a task only if
the directory exists - (fopen,fwrite - create task check file)

Will that be more beneficial? or how will I know it's too much/many?

Part 2: QUESTION

I've got a second routine to loop through and find 4 'jobs' tasks & update
the database via MYSQL.    If I improve the 1st part (as above) there could
be a potential of 100's or 1,000's of job tasks to do...

currently, ive got a loop to process 4 'jobs' tasks at once

- find job tasks
- Get raw data
- update database SQL.
- delete task file
- sleep(1)
- repeat 4 times.


If I remove the sleep(1), could I process 100 at a time? More? - without
breaking anything?