Duration of loading tickets_incremental can vary widely, depending on whether a full (re)load is done or only an increment. On a full (re)load - even when just 1 row is retrieved - the whole dataset will be retrieved.
Typical 95% probability response time for full (re)load is like 1 second per 100 rows in the full dataset. For instance, when there are 100.000 tickets, the initial response time can be as much as 1.000 seconds, even for 1 row.
Once the whole dataset is cached, the duration of an update is significantly shorter, and can be as little as 15 seconds for the same datavolume depending on the available CPU capacity and network latency.
The actual API calls on Freshdesk can be found in the form “Session I/Os” (archived every day at 00:00 UTC).
Can you include an (anonymized) screenshot of the associated request from Invantive Bridge Online Monitoring (see More insight with new Bridge Online Monitoring) with the title including request ID visible?