Delay to load API response | Bonitasoft Community

API 1: sessionUser (../API/system/session/unusedid)
API 2: humanTask (../API/bpm/humanTask?c=-1&p=0&f=user_id={{sessionUser.user_id}})

In my application page I use a task filter for each user of the session, displaying the result in a table.

I noticed that when I run the config API calls described above, the task API response time is high (2-3 seconds). Being that, I tested it on an application that only performs this.

How can I make this type of filter to perform better with a faster return?

#api

Log in or register to post comments
1525 reads

Comments

Submitted by antoine.mottier on Mon, 04/08/2019 - 15:57

First of all I would recommend to avoid using -1 as result count for pagination. You might want to use a large number but having no limit is always a risk.

I did a few very quick tests and didn't manage to reproduce a response time of 2-3 seconds. How many results did you get from the API call?

Permalink

Submitted by jhonata.v.t_1373703 on Wed, 05/15/2019 - 20:55

The task call returns approximately 200 tasks and after the user filter returns up to 20 tasks.

The tests done on localhost are fast, when inserting the page into an application the flames return from the APIs become slow.

What factors may be causing these delays? Number of pages in the same application? Number of API calls executed at the same time? Number of forms inserted in the Portal executed within the page? Do you have these performance data?

Permalink

Submitted by antoine.mottier on Thu, 05/16/2019 - 10:05

You can find some information about performance troubleshooting in the documentation.

Number of pages in an application should not impact performance.
Number of API calls executed at the same time can impact the performance depending on the complexity of the processing they triggered and the volume of data they produce.
The number of widgets and variables declared in a page can influence performance.

Maybe you can share your project export from the Studio as a .bos file and I can try to reproduce your performance issue and isolate the root cause?

Permalink