parallel request problem

Hi there,

for a stress test we try to run a workflow on the knime server, that should only return a dummy json response, from a client with a post request and a parallel chunk loop.

If we split 100 requests into 2 parallel chunks we get a 200 http response each time, but the json-output field does not exist.

So, why should the knime server respond while the workflow isn’t finished with execution?

Best regards

Thanks for the question. How do you submit your requests to the KNIME Server? If you are able to share the workflow here then I can take a look at it for you.

Best,

Jon

Hi Jon,

thanks for the answer. It’s just a json reader that opens a dummy file an put it into the json output.

And the stresstest caller.

local_client.knwf (12.6 KB)

fast_running.knwf (9.1 KB)

Best regards

I ran the workflows that you sent to me after adapting them to my KNIME Server address.

I made sure to increase the timeout on the POST request node since I was making the calls to a remote server on the internet.

I tried with the POST request node inside and outside of the parallel chunk loop and there wasn’t a difference in the response (always a 200, and always I could extract the name ‘John Dow’ from the response.

If you’re still running into issues, perhaps we should setup a call with one of our support engineers to look into the issue in a bit more detail.

Best,

Jon