Knack API Performance

Does anyone else experience The knack API being very slow compared to other APIs?
Which ever type of request (e.g. search, update, get record) I perform it is extremely slow, When I call other API to update Knack the user has to wait a while which is not ideal. Simarly when I call the Knack API from external systems.
This happens whether I use the knack API via Integromat or locally via postman.
I am happy to pay for a more performant API but there doesnt appear to be an option.
I am interested to hear others experiences with the API and what tips you may have in getting a more performant API

Yes I too face this issue when I am executing through 'make".

I dont think that this is limited to make though, I get the same performance when interacting with it directly. One thing that i have to have an error handler on every knack module though is for “503: upstream connect error or disconnect/reset before headers. reset reason: connection termination”

Yes Amaan I too handle the same way using ‘ignore’ on error handler. It dalays make scenerio(complex) execution in my case which is annoying for my users.

You would probably want to use Break instead so that it retries it?

Yes! I’m getting lots of 429: Rate limit exceeded also…

I moved our automation from Zapier to Make in September 22 because I was having problems with the new Knack upgrade. Things seemed to go pretty well until recently when about 2x week the Make automation stops executing because of a 429 rate limit exceeded error. The automation looks for an updated record, then evaluates a couple binary fields and then depending on those fields, sends a text message through Text Magic. Nothing fancy going on. I have added (2) delays steps each set for 15 seconds and it still hangs up. Should I shorten the delays? Those were #s given to me by Make Support. I’m wondering if the delay is simply too long.

Hi @Scott3 you might be better off without the delays but use the error handler of Break and set retry to 1 minute, especially if it only happens twice a week. That way the rest of the time it’s responsive


Thanks for the response. I changed the delay from 15 seconds to 3 seconds on 1/16 and so far I haven’t received any errors. This is the longest period since December the automation has run without intervention.

I did go in and add a Break today but left the delay in.

My guess is the 15 sec. delay was working against itself. This automation looks for an updated record and if several records were being updated, when the records processed in Make a log jam occurred somewhere. Of course I don’t have any way to track this other than look at the records that fail and then compare them to the update logs in Knack but that takes a great deal of time to compare the individual logs.

The Make team suggested the long delay, but when you think about it anything longer than 5 seconds is an eternity today.

For now I will leave the 3 second delay in place and use the Break as a way to handle any errors that do come up.

Thanks again,


1 Like