ssx

app syncing

Recommended Posts

ssx    2

I have been using inoreader with mr reader. Unfortunately, the past few days, not all the items in my inoreader are syncing with the app. I contacted mr reader and was told to contact you to inform you that the latest api changes you all have made are buggy which is why all my articles won't sync

Share this post


Link to post
Share on other sites

Hello, 

Could you please give us little more details about the issue?

What do you mean not all items are synced?

Please note, that there is limitation for maximum items to return, which is 1000. And I think that this this is the issue you are facing.

You can see our API documentation where we explain this.

So there is nothing buggy here ;).

Share this post


Link to post
Share on other sites

Roughly how many unread articles do you have? Do you experience sync issues across all folders/feeds or only the ones with the most unread articles? The limit was always officially 1000 articles, but it was actually set to 10000. Lately we had a lot of headaches with this, because this means that 10k articles should be loaded up and held in memory during the request, their tags, read flags and other attributes should be checked and a also recorded in-memory and then spit out to the client. When enough such requests are being simultaneously processed there's a high risk of servers running out of memory. There is a memory limitation per process to minimize this risk (but only to extend because the maximum number of processes is always pre-provisioned, so max_processes*max_memory_per_process is always > available memory) and some users were facing sync issues (complete sync failures), because their processes were reaching this limit. We have provided a means for paginating the results, which can break those huge requests into smaller chunks (with the continuation parameter), but Mr. Reader doesn't use it and makes one big request for 10000 articles instead. If somehow you persuade the developer to break this huge request into smaller ones, the issue will be resolved.

Share this post


Link to post
Share on other sites
ssx    2

This is absolutely awful because I cannot use mr. reader with only 1000 items - I always download the maximum number of unread items. I will email mr. reader to let him know what you said

Share this post


Link to post
Share on other sites

I have the same situation with Reeder (described in the parallel thread). 

It is indeed absolutely awful, but I'll let the developers know.

Sounds a bit like a chicken egg problem: 1000 unread articles are not really much these days with a decent number of feeds. Also memory (in VMs), memory management, request scalability and database backend shouldn't be that vulnerable to such requests these days. Nevertheless programmers should of course use the proper API calls to send proper requests if they are implemented in the API. 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now