Webhook (APP) throttling doesn't work?

[ralphioooo]

I've created an APP which uses a webhook but having issues with too many requests coming through at one time:

https://developers.hubspot.com/docs/api/webhooks

I've played around with the throttle settings. But even when I set it to something like 20 per minute I my endpoint gets hit at warp speed. 

Eg. Just tested 2000 contact updates on 1 property. Throttle set to 20 per min. (waited at least 5 min after setting this)

When it ran I got about 35 requets over the space of 5-10 sec. 

Tried quite a few different combos (waited at least 5 min) but made no difference with delivery speed.

[WendyGoh · Accepted solution]

Hey @ralphioooo,

That is absolutely spot on. 

In the scenario you shared, since per minute you're set to receive 20 requests which is equivalent to 2000 events 

The first minute you'll receive 2000 events and the next minute you'll receive the remaining 300 events. 

Does this make sense?

[WendyGoh]

Hey @ralphioooo,

Just to clarify, what's the rate limiting that your team set on your system? 

Additionally, I'd like to clarify the following:

The event throttling, applies to request not to event. So if you’ve a system where it takes you like 5 minutes to process one single event and you put your throttling down to 5 requests per second. If you have 500 notifications, you get 5 requests within 1 seconds that has a 100 notifications in each requests. So that 500 events that you’ve to process at your own speed.

[ralphioooo]

Hey Wendy!

I tested quite a few different combos and didn't seem to make any difference. (perhaps my contact set isn't large enough?) 

Currently have it set to 20 per minute. When I bulk updated lead score for 2300 contacts I can see around 30-40 seperate request containing varying number of events per request. All coming at once/within 10 seconds. 

In this scenario would the first 2000 not come through in the first minute with the remaining 300 a min after? (assuming 100 per request so 20 x 100 ?)

[WendyGoh · Accepted solution]

Hey @ralphioooo,

That is absolutely spot on. 

In the scenario you shared, since per minute you're set to receive 20 requests which is equivalent to 2000 events 

The first minute you'll receive 2000 events and the next minute you'll receive the remaining 300 events. 

Does this make sense?

[ralphioooo · Accepted solution]

Thanks Wendy, Also been chatting on the dev slack channel. Found out the issue with the errors. Thought I would add it here for anyone else who is getting confused by it:

"The first retry happens pretty quickly, within seconds. Looks like in your example, the first delivery failed because it hit the concurrency limit, thus it couldn’t deliver and had to retry"

I was getting concerend as they were coming up as errors and thought it was failing to delivery but sounds like it's actually the throttle doing its job 🙂

Thanks for your help.