How to Setup current database for accurate asynchronous pagination - sql

Question

So currently, I am working on a way for accurate pagination for an asynchronously written database. Basically a request comes in and is broken down into X number of Jobs (in a Queue) which get pulled by another app which batch writes Y number of ResponseItems per X job to a PostgreSQL table.
Two Tables
RequestMetaData: RequestId, ..., ..., ...,
ResponseItemData: Id, RequestId, ..., ...
Each ResponseItemData is part of a Request.
Lets say a request A comes in. It gets split into 700 jobs. Application pulls a job and writes 1000 items. But another request B could have come in as well, so the next job that gets written could have written 1000 items.
1
.
.
.
1000 ==> Last Item in Job 1 of Request A
1001
.
.
.
2000 ==> Last Item in Job 1 of Request B
2001
.
.
.
3000 ==> Last Item in Job 2 Of Request A
So my question is how do I do accurate pagination here? I already know how to do relative cursors
and can retrieve X number of items per requestId by using the last item seen in a incrementing table. Something like this
SELECT *
FROM `ResponseItemData`
WHERE `ResponseItemData`.`id` > **$Last_Item_Seen(Null On First Get Request) ** AND 'ResponseItemData'.'id' == **$RequestId**
ORDER BY `ResponseItemData`.`id` ASC
LIMIT X
My question:
A) How do I know when a request is done? If I keep paginating using a requestId 1000 items 699 times how do I know the 700 is the last one? How do I keep track of when my request is done? In my RequestMetaData Table, I store the amount_of_jobs_request_split_into as well as the expected_number_of_response_items.
Thanks for the help