Is the API going to break every weekend?

[witcher.3197]

I don't mean to be disrespectful but a lot of GW2 sites use the API, and it's seemingly the second week in a row where it dies on Friday and stays off for the entire weekend. A lot of people only have time to play on weekends and they are repeatedly denied these services..

[witcher.3197]

Oh look it's dead again..

[Killerassel.2197]

For the record: the API itself seems to be running fine. The rendering service, i.e. icons for items and skills, throws errors. And it didn't just start to do that today.

It would help if the other sites cached the icons locally and would not entirely depend on them to convey information.

[DragonHeart.7201]

@Killerassel.2197 said:For the record: the API itself seems to be running fine. The rendering service, i.e. icons for items and skills, throws errors. And it didn't just start to do that today.

It would help if the other sites cached the icons locally and would not entirely depend on them to convey information.

Actually almost every individual endpoint had drops weekend, just seems the rendering service is the one most affected by the bug or whatever it is. So even though it mainly is the rendering service being a pain, it isn't exclusive to it.

Getting a bit sad there has been no response at all on any of the threads made, even though there still is a developer working on the API.

[Broadway.8925]

I'm with Witcher on this one. I'm sick of this crap. This game seems to be having one issue after another. Staff layoffs, little to no attention to World Versus World (despite the game boasting world combat and the name being Guild Wars), bug fixes being a regular occurrence after every patch, and API issues being the icing on the cake. I am a pay-to-play'er. I spend too much hard-earned money on this game to not be able to look up my saved build templates due to the instability and chaos. I've bitten my tongue for weeks now, but I can't handle it anymore. If ANET doesn't get it together soon, I'll take my business elsewhere and watch as the game continues to plummet and most likely die. I've seen this happen with MMO after MMO. It's not pretty, but it happens.

[TwiceCookedPorkins.8940]

This is getting ridiculous. EVERY. SINGLE. WEEKEND.

[StevenL.3761]

Friends, let's assume nobody is doing this on purpose. Whatever the reason is for all the outages, yelling at the problem isn't going to fix it. We can do better if we post timestamps and details of requests and responses so that the devs have all the information they need to diagnose the issues.

[Broadway.8925]

Details and timestamps of what? It's not an in-game issue. All of the sites that use the API for anything are essentially useless at this time, and that's making playing this game Hell for some of us. Someone just needs to fix the API and try to keep this from happening SO OFTEN (because, I know it's impossible to prevent) again.

[DragonHeart.7201]

@"StevenL.3761" said:Friends, let's assume nobody is doing this on purpose. Whatever the reason is for all the outages, yelling at the problem isn't going to fix it. We can do better if we post timestamps and details of requests and responses so that the devs have all the information they need to diagnose the issues.

Ofcourse it's not happening on purpose, no one is implying this. Though there have been multiple threads explaining what errors are happening and when, there is plenty of info out there, without any response from the dev working on the API.

Will add some info again:

Request URL: https://render.guildwars2.com/file/A66CEF1B9C032C240EDB9077FCD53493CD5AC6B9/1203050.pngRequest Method: GETStatus Code: 500Remote Address: 13.33.91.225:443

Response:

{"error":48,"product":0,"module":2,"line":89,"text":"file couldn't be loaded"}

Timestamp:date: Mon, 03 Jun 2019 10:36:42 GMT

[Broadway.8925]

I heard that there is currently no API dev at all. I think I'm going to sideline GW2 for the time being. It may just be time to let the game go and move on to something else.

[Robbes.2317]

@Killerassel.2197 said:For the record: the API itself seems to be running fine. The rendering service, i.e. icons for items and skills, throws errors. And it didn't just start to do that today.

It would help if the other sites cached the icons locally and would not entirely depend on them to convey information.

it's not an API fault, it's the design flaw of sites that are using it, if they cached the things they request, they wouldn't need to rely on the API evrytime someone refreshes their site. those sites choose the easy way out, which only taxes the API and not their own site. it's just overwhelmed.

[Illconceived Was Na.9781]

https://en-forum.guildwars2.com/discussion/78590/render-server-outage@Daniel Snider.6241 said:Hello,

I know you all have noticed the API breaking frequently recently and I wanted to write a post about it to help describe exactly what parts are breaking for clarity.

render.guildwars2.com is a piece of the API that converts game asset images into jpgs and pngs for the public to use in their apps and sites. It's backed by a pretty simple standalone piece of server software whose sole responsibility is to cache the assets, convert them, and serve them as images.

This server has been getting stuck in a bad state about every other month or so for a while now (I think a little over a year). It has been on my backlog of things to fix, but since the solution is a gentle reboot now-and-then, I admit I flagged it as low priority.

Over the last few weeks, for reasons I haven't yet figured out, it's been getting stuck very frequently, sometimes twice in a day and at least once every weekend. It's gotten so bad that due to server configuration reasons, it wiped the white list server which had to be recovered manually (which is why the item endpoint died recently). Thank you for your reports of failure on the forum. I have been using these reports to manually fix the server while we try to work on a real solution.

The priority for fixing render.guildwars2.com has been bumped up and is being actively worked on. The current status is debugging and gathering info on why it's breaking. As soon as we know, we'll fix it as soon as possible.

Thanks again for your reports, and thank you for your patience while we try to put out this fire for good.

-Snider

(I pasted the response here because at least one poster above seemed to have missed it.)

[Healy Unit.1023]

@Robbes.2317 said:

@"Killerassel.2197" said:For the record: the API itself seems to be running fine. The rendering service, i.e. icons for items and skills, throws errors. And it didn't just start to do that today.

It would help if the other sites cached the icons locally and would not entirely depend on them to convey information.

it's not an API fault, it's the design flaw of sites that are using it, if they cached the things they request, they wouldn't need to rely on the API evrytime someone refreshes their site. those sites choose the easy way out, which only taxes the API and not their own site. it's just overwhelmed.

No offense, but you clearly have no idea what you're talking about. If the API were returning 429s (rate limited), 500-level errors, or something else, (or simply timing out), then yes it'd likely be an overload issue. However, it's returning a 404 (not found) plus a custom status message, which is not indicative of a server being "overwhelmed".Furthermore, sites like gw2spidy actually do cache their data somewhat, as they use it for graphing price-over-time data. However, for a site like gw2spidy, relying solely on cached data is actually impossible; they'd not be able to show updated (i.e., current) price data at all if they did that.

[Stephane Lo Presti.7258]

Hi everyone,

As mentioned above, @Daniel Snider.6241 provided an update about the API availability and the work that he's doing to fix it. I'm going to step in and close this thread based on the aggressive tone of initial posts and the fact that we want the API subforum to focus on communication to and from the API-based app developer community.

Thanks for your understanding