You guys are screwing yourselves with sloppy infrastructure setup (API problems)

[DankTemplar]

I was going on a little bit of a download spree for Skyrim LE, got to around 3/4 of the 400 mods I needed to download when I was spammed with that fun API limit, which I get, if you didn't have that, you could end up with some major issues.

 

My problem is that all the queue'd downloads have now error'd out, meaning I'm going to have to go and spam APIs again after figuring out what mods haven't downloaded which is near impossible.

 

Guys, I know money is probably tight for Nexus but you've got to step up your game. This creates more load on your servers than it would have had if your systems worked properly.

 

How can you expect me to pay for a service when your limitations to prevent your servers from getting overloaded are so ham fisted that downloading 300 mods triggers it and then instead of having downloads paused, the server does something as stupid as just cuts off the connection period?

 

You really, really, really need to get it together when it comes to the fundamentals like this.

[Arthmoor]

Actually it sounds to me like the safeguards in place kicked in and did exactly what they're intended to do in your case.

[Deleted67195711User]

But DarkTemplar does make a valid point. Had the safeguards actions simply requed the download(s) for later instead of flushing the queue, it would have saved on the API activity necessary for the user to find and restart all the unfinished downloads.

So yeah, the safeguards worked. But they did not necessarily cut down on the API activity and may have acted in such a way as to actually increase the load.

[Ethreon]

"How can you expect me to pay for a service" - which service are you paying for again?

[BigBizkit]

Thanks for your feedback. This article has more information about our rate limiting: https://help.nexusmods.com/article/105-i-have-reached-a-daily-or-hourly-limit-api-requests-have-been-consumed-rate-limit-exceeded-what-does-this-mean

 

It seems the necessity for having limits is clear to you. That being said, we stated before that we maintain an open mind when it comes to determining the limits (i.e. if they need to be higher, we might consider it). However, so far we have not received many complaints regarding the current limits. Your feedback is noted, however.

[Reneer]

"How can you expect me to pay for a service" - which service are you paying for again?

This confused me too. Unless they have another account, it doesn't look like they paid for premium (maybe it expired)? Or perhaps they are using NXM links?

 

As for queuing / resuming API downloads, it's really a client-side thing for the most part (assuming everything is going over at least HTTP 1.1) since the client needs to support range headers (though I'm assuming here that the servers do as well) and the file operations necessary to restart the file download at X bytes. All the API is doing is giving you a download link - it is the responsibility of the client to handle the download properly and keep track of it.

 

The HTTP range header is designed to make the client responsible for resuming downloads rather than the server (as it should be), since there are only one or a few servers and many clients. Making the server responsible for this would just bog them down needlessly.