Does the API for Wvw timers and kills have an issue since a few weeks? — Guild Wars 2 Forums
Home API Development

Does the API for Wvw timers and kills have an issue since a few weeks?

Nauda.3678Nauda.3678 Member ✭✭
edited November 22, 2017 in API Development

Not sure if this is the right place but I think it might have to do with the API.

I'm using a lot of sites like:
https://wvwintel.com/
http://kills.werdes.net/
http://gw2timer.com/?page=WvW

And I noticed that objects often don't get updatet properly. In some worst cases some captures are not being displayed for hours.
It's also very noticeable for site that track kills (like this: http://kills.werdes.net/?match=518 ). You can see how there are some hugh gaps or low amount of kills in the graphs that don't make sense. Sine it seems to affect different site I assume that it might be a problem with the API.

This week http://kills.werdes.net/?match=459 (2017.10.06 - 2017.10.13) seems to show the start of the gaps. Starting on tuesday the gaps get more and more prominent. The week after it looks like this: http://kills.werdes.net/?match=468

So I just wanted to know if anyone knows about these problems and if they maybe get fixed?

Comments

  • Yes, I'm seeing the same thing. Frozen data.
    You can test it directly. Fetch https://api.guildwars2.com/v2/wvw/matches/1-3. Wait 10 minutes. Fetch it again. All the data is the same.

    If you have linux or a mac, you can use
    curl https://api.guildwars2.com/v2/wvw/matches/1-3 | md5
    To generate an MD5 fingerprint each time you fetch.
    Wait ten minutes then do it again. Same MD5.

    But every once and a while the API will start reporting again.

    The API just started again as I was editing this. It started at a 2 hour skirmish reset. I expect it will freeze up after 20 minutes again as it did yesterday.

  • I made a bash script to monitor the updates of www/matches/1-3. It's going up and down. I don't see a pattern. Since the sampling was down once a minute it is possible the short durations of no change may happen in the early hours. Looks like a nasty bug.

    Updating: Thu, 23 Nov 2017 18:10:43 -0700
    Not updating: Thu, 23 Nov 2017 18:50:33 -0700
    Updating: Thu, 23 Nov 2017 19:48:56 -0700
    Not updating: Thu, 23 Nov 2017 19:49:57 -0700
    Updating: Thu, 23 Nov 2017 20:49:35 -0700
    Not updating: Thu, 23 Nov 2017 21:56:40 -0700
    Updating: Thu, 23 Nov 2017 22:41:58 -0700
    Not updating: Thu, 23 Nov 2017 22:44:00 -0700
    Updating: Thu, 23 Nov 2017 22:55:18 -0700
    Not updating: Thu, 23 Nov 2017 23:59:44 -0700
    Updating: Fri, 24 Nov 2017 00:58:10 -0700
    Not updating: Fri, 24 Nov 2017 03:00:27 -0700
    Updating: Fri, 24 Nov 2017 03:33:29 -0700
    Not updating: Fri, 24 Nov 2017 03:35:32 -0700
    Updating: Fri, 24 Nov 2017 04:00:04 -0700
    Not updating: Fri, 24 Nov 2017 04:07:11 -0700
    Updating: Fri, 24 Nov 2017 04:33:41 -0700
    Not updating: Fri, 24 Nov 2017 04:35:43 -0700
    Updating: Fri, 24 Nov 2017 04:44:47 -0700
    Not updating: Fri, 24 Nov 2017 04:46:50 -0700
    Updating: Fri, 24 Nov 2017 05:06:17 -0700
    Not updating: Fri, 24 Nov 2017 05:14:21 -0700
    Updating: Fri, 24 Nov 2017 05:21:22 -0700
    Not updating: Fri, 24 Nov 2017 05:23:25 -0700
    Updating: Fri, 24 Nov 2017 05:40:37 -0700
    Not updating: Fri, 24 Nov 2017 05:42:39 -0700
    Updating: Fri, 24 Nov 2017 05:54:47 -0700
    Not updating: Fri, 24 Nov 2017 06:00:47 -0700
    Updating: Fri, 24 Nov 2017 06:01:49 -0700
    Not updating: Fri, 24 Nov 2017 07:10:45 -0700
    Updating: Fri, 24 Nov 2017 08:10:20 -0700
    
  • don't know if it's related, but in the past 24 hours I've seen an increase in the 206 partial content responses when querying endpoints with lists of ids just retrieved from those endpoints.

    I do think there's something up with the APIs caching layer

    Black Lion - Roar on the Street : http://www.gw2roar.com/

  • atheria.2837atheria.2837 Member ✭✭✭

    @Nauda.3678 said:
    Not sure if this is the right place but I think it might have to do with the API.

    I'm using a lot of sites like:
    https://wvwintel.com/
    http://kills.werdes.net/
    http://gw2timer.com/?page=WvW

    And I noticed that objects often don't get updatet properly. In some worst cases some captures are not being displayed for hours.
    It's also very noticeable for site that track kills (like this: http://kills.werdes.net/?match=518 ). You can see how there are some hugh gaps or low amount of kills in the graphs that don't make sense. Sine it seems to affect different site I assume that it might be a problem with the API.

    This week http://kills.werdes.net/?match=459 (2017.10.06 - 2017.10.13) seems to show the start of the gaps. Starting on tuesday the gaps get more and more prominent. The week after it looks like this: http://kills.werdes.net/?match=468

    So I just wanted to know if anyone knows about these problems and if they maybe get fixed?

    It seems after the last update, everyone in Anet left and decided the game didn't need any more maintenance or monitoring.

  • Still seems to be a problem sadly. It greatly decreases the use of wvw sites sadly:(

  • Nauda.3678Nauda.3678 Member ✭✭
    edited December 21, 2017

    Still seems to be a problem sadly. It greatly decreases the usefulness of sites for wvw object timers :(