01-16-2018 01:12 AM
01-16-2018 01:12 AM
Starting last Friday we started to receive an significant increased amount of 429 responses from requests made with CSOM from our SharePoint Provider Hosted Add-in. The same started to occur on multiple tenants at the same time and is now a problem on up to 60 different tenants. We are using best practices to avoid Throttling like decorating request with a User Agent and are using Office Dev PnP ExecuteQueryRetry for executing all requests. Before Friday this haven’t been a problem on any tenants. No changes are made from our site to cause this. Any help would be appreciated to shed some light on this issue?
02-19-2018 07:56 AM
I do apology for a long delay in the response, but let's get moving with this one. If this is still a problem, please do let me know using private message (email or TechCommunity) and we can start a discussion on the specifics. I'll then loop in the actual owners of the throttling capability.
02-21-2018 08:38 AM
Thank you for taking your time to consider this issue
This is still an issue for us and for other ISVs as well. We did raise a ticked that ended up concluding that our multi-tenant add-in was categorized as an app that performs bulk operations. It has now been classified as interactive traffic and the problem decreased for the majority of our customers.
We do understand the need for throttling. But before this issue started we never received a single 429 response from SharePoint. Now, on some tenants we receive a very high percentage 429 responses compared to requests. For example we have one tenant where we in a 4 hours timeframe had 260 request to SharePoint where 83 of them is Throttled.
Most worrying is it that in periods where tenants are Throttled we cannot rely on CSOM ExecuteQuery requests. We often see that we receive 200 responses even that the request has not loaded the properties requested. This of cause making it very difficult developing business critical solutions on SharePoint.
How SharePoint online uses throttling is not very clear to us. Our application is rather complex doing light weight stuff against SharePoint in remote event receivers in a user context. On the same application we create sites on request in web jobs, synchronizing list and security to the sites executed in an app only context. It seems that SharePoint does not distinct and both cases are throttled equally. The last one we can handle with exponentially back off. The first one we can not…
Any help on how to handle this is very much appreciated...
02-21-2018 09:58 AM
this is not really around us considering if this is an issue, it's clearly an issue for some of our ISVs, so we are trying to get the details and contacts to have an open one-to-one discussion on the right patters to move forward.
Can you share your email (for example private message) or send me directly an email with ISV solution details, so that we can move on this more efficiently. Thx.