Problem

When REST call to extract some pages from Confluence is called, if the pagination start page is more than 50k, the system takes close to 8-9 seconds to return back with the response. This response is faster if the start page is lesser. 

Environment

.Confluence 6.13.8

Steps to Reproduce

1. Create about 70K pages in Confluence under a single space. 
2. Once created, use the following curl command to extract pages, starting from the 50,001

curl  -u username:password -X GET "http://localhost:port/confluence/rest/api/space/SPACE_KEY/content/page?limit=20&start=60000

3. The command takes about 8-9 seconds to return with a response.
4. Try the same command with the start value with about 10000. The pages are returned much more quickly

Expected Results

Both the calls for pages starting from 50k and pages starting from 5k should take similar times.

Actual Results

There is considerable difference between when pages are fetched with start of 5k to that of 50k.

For 5k start, the call takes about 0.8 seconds :

time_connect: 0.001467s
time_appconnect: 0.000000s
time_pretransfer: 0.001512s
time_redirect: 0.000000s
time_starttransfer: 0.873332s
----------
time_total: 0.889090s

For a start of 50k, the page takes about 9 seconds :

time_connect: 0.001452s
time_appconnect: 0.000000s
time_pretransfer: 0.001492s
time_redirect: 0.000000s
time_starttransfer: 8.886917s
----------
time_total: 8.971300s

Workaround

No work around at this point.

Notes