the intermittent failures should resolve themselves as some services come back online
@Nabrok hmm, so I didn't think about this, but the kill/death counters are accumulated in-memory, so when the backend servers get restarted (e.g., due to hotfix) they're purged.
I need to serialize out those values I think.
It's super annoying but the aggregate scores from
/v2/wvw/matches should be correct, though the ppt/ppk breakdown will probably be wrong for awhile :(
(maybe s/correctly//)
haven't figured out how but it's causing problems
here's the number of bytes retrieved from the cache/time: http://i.imgur.com/2sfIluB.png
gonna sleep on it but definitely open to suggestions. problem is mostly that the responses (which are cached wholly) are too large and sucking up too much space.
Also @BryghtShadow you definitely should! You seemed to be a good editor from the changes you suggested to me, and API editorship isn't hard to get.
I don't think anybody else does that
thats what i wonder when i read all those qq threads when a server comes down for maintenance
but indeed, the very low downtime is noteworthy about gw/gw2
but that has been changed fortunately
I thought it was only on
/v2/account
I may have mis-merged something
just updated the file
and just to clarify, "commander" will be true if they have purchased a tag?
(not if they are tagged up)
could (eventually) expose party/squad status but not sure if worth
(and how the squad is organized)