let's see, is there any custom parameter missing here that anyone know of?
- /v2/commerce/exchange (gems, coins)
- /v2/guild/:id/log (since)
- /v2/pvp/seasons/:id/leaderboards (ids)
- /v2/recipes/search (input, output)
- /v2/wvw/matches (world)
- /v2/wvw/matches/overview (world)
- /v2/wvw/matches/scores (world)
- /v2/wvw/matches/stats (world)
this is what @queicherius's library has in the docs on my branch atm
- /v2/guild/search (name)
You going through a documentation journey? :angel:
well, i updated the docs to reflect my code changes :)
hmm @darthmaim, indeed, i missed that one since the doc structure is different for that one
Ah ok cool! Maybe I should move over to use gw2api lib :-)
argh, @queicherius doesn't use
;
if everyone could gather round real quick, I'd like us to gather to share a few moments of silence for perforce
:pray:
:fire:
:pray:
stupid me,
:fire: :pray: :fire:
:pray:
:pray:
well i'm going to be here all night
i've got a pending submit which turns the API back on
but the p4 server is hung
:(
and by "pending" i mean "uncancelable in-flight"
lye @lye cackles madly
@queicherius i created the PR at queicherius/gw2api-client#23 and my test failure at queicherius/abc-environment#35
@Archomeda Thanks <3 I'll have a look at the environment thing tomorrow and at the client when the API is back up. :+1:
i'm just baffled at
npm test... i mean the files are there
I have no idea either, will have to test.
but yeah, whenever you have time :smile:
Does the windows bash do shell expansion?
@queicherius msysgit's bash shell does
i'm not that familiar with bash at all, what kind of expansion do you mean in this context?
I was guess globbing, e.g.
echo *.c ???bar
cat {src,tests,.}/**/*.js
yeah that
oh, i should probably note i'm running this in powershell
my best guess, probably not
Is
** a thing?
powershell lols
hey, better than cmd
and webstorm doesn't properly support bash in windows
"stabbing yourself in the foot is better than stabbing yourself in the hand"
you're not wrong
but you're also not right
IT FINISHED
:tada:
Wool
-l
grats! :D
now we gotta wait for the configs to propagate
I always use
** / globstar for recursive directories, dont know if I picked that up wrong
No that is correct
huh new features for me
- is only one level, ** is all subdirectories
I normally use regular old borne shell, so y'all and your newfangled features
i've seen this in many libraries that interact with the file system
not only bash
hell, i think even windows supports it
(don't quote me on that)
"hell, i think even windows supports it" -- Archomeda, 2017
good, not me then
""hell, i think windows supports it fully" -- Archomedia, 2017" -- lye, 2017
well crap
chinese whispers?
lol
that edit is not gonna save you :whistles:
okay I'm seeing traffic
it's routing to the correct rate-limiting servers, yay
:fire:
madou shhhhhh
https://api.guildwars2.com/v2 IT LIVES!
@queicherius gw2e IPs should be whitelisted so you don't have a rate limit
everyone else -- you're stuck on 60 reqs/min for a bit
:')
thanks obama
damn elitists
pug life... wait
also
actually I'll just slam it myself
yay, my discord bot went crazy and posted the guild log backlog
I want to test something quick, no :fire: yet plz
@lye Thanks <3 All backend things should currently be disabled, so if stuff breaks it's not me!
sadly most of the entries fell off the log :(
dang
rip, fire
oh well
okay managed to trigger the rate limit
so that's working
:tada:
apparently chrome is not fast enough to hit it with constant ctrl-shift-rs
i totally hit it with chrome
it's once/second for a minute
also why are my graphs broken I am so confused
I tried hitting
/v2/gliders a bunch D:
huh, got the same behavior
i think chrome aborts the in-flight requests if you click f5 too fast
umm
hmm
why not run wget in a loop if you're a unix user anyway ? :P
this is interesting
/v2/build gives me the error but /v2/gliders doesn't
I knew it
🤢
gonna be a long night
don't do that to yourself lye :worried:
i mean
i'd rather spend an hour now than get a phone call in two
eh.. hm
and my phone call I mean my housemate gets the phone call and then knocks on my door to tell me he got a phone call
and say a bunch of other things?
like "did you light the server on :fire: again?"
gw2@naga:~$ for run in {1..120}; do curl --write-out %{http_code} --silent --output /dev/null https://api.guildwars2.com/v2/build; done
200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200200
just don't forget the time and realize you're still in the office at midnight :(
it looked like it was averaging about 3/s
Ey, well api was up long enough for me to debug my code snipet <3
did you want more pings to the API?
I'm trying my best to hit the limit
i'm so confused
I added
throw 'ratelimit' to the ratelimit code
/v2/gliders is the only endpoint i've found that doesn't hit it
also my local copy is successfully logging per-endpoint requests?
I'm really not sure how this broke
it's the newest one, so maybe something funny is happening with new stuff?
4 consoles running a curl each couldn't hit limit on gliders
@Eearslya gliders doesn't apply ratelimiting for reasons that I cannot discern
like, it does not execute the ratelimit code
I don't know why
I don't know how
I'm pushing a patch through now
magic switch not flipping?
gremlins
literally gremlins
@Darqam haha, i noticed @ChieftainAlex requested permissions for me; that response though :D
do you hate gremlins?
well, at least it's working on
/v2/items right? that was the biggest culprit before
I'm testing on several endpoints right now
have to sleep between tests though
items does give 429s
it works on
/v2/items, tested that locally
it's a lot easier when your local config is 6 reqs/min
tested build, gliders, items, recipes, wvw/matches...yeah, only gliders failed
it makes no sense
I implemented ratelimiting by modifying the
router.get bit
so that it invokes the ratelimit check before the controller
is there a hardcoded bit in the ratelimit backend that matches the glider endpoint? :worried:
and skips it?
but in
/v2/gliders (which uses the same routing bits as everything else in the same one file) doens't execute it
yeah I put a
throw statement in the ratelimit code just to sanity check
and it literally doesn't hit it
took me forever to dig this up, but Lye this might explain why things are bad (it's fluff, no it wont): Gnomes run your computer
Is IIS hijacking the request somehow?
shouldn't be
since I don't have IIS running locally
and it still repros
welp that's all I had
gremlins
there's probably a totally logical reason behind it, but figuring it out would take 2h
so I just rejiggered how the ratelimiter hooks in
add
/v2/gremlins and return {text: "everywhere"}
pretty much
Gremlins. Litteral Gremlins.
p4 could have chosen a much better day to throw temper tantrums
dwayna save me
well
I don't think I can actually do anything else tonight
the api isn't gonna fall over? :smile:
i mean
i wouldn't bet on it
alright
still 60/m?
yeah
fair ^^
i'll up it once i get this fix to
/v2/gliders
need to be able to verify the fix on live
appreciate what you're doing <3
i'm really upset that this was down for so long, so
have a good sleep though :)
going to pound some tequila and get a couple hours of path of exile in
that should do the trick
:D
:fire: doing a live deploy :fire:
I really hope this magically fixes the perfs
nope
but hard-restarting the processes does lol
WELL THEN
wut
dude i don't even know
there's so much going wrong today I don't even know where to start
Perf gainz
okay, verified that
/v2/gliders is now magically rate limited
for reasons that I do not understand.
:sparkles:
going to up the rate limit to 600 reqs/minute
@Eearslya nice find on that btw
I started testing
/v2/items but the request was too big, loading too slow to hit limit, so I just went with a smaller endpoint. /v2/buildnever occurred to me XD
lol
well thank goodness for that
Thanks for your hard work lye
seems like it's working
@madou no problem, thanks for bearing with me
5k requests hit the limit when 60/min? :O
hahahaha yeah apparently
lord only knows
probably still that
/v2/items...
Re: the armory embeds, I hope after I put a few more smarts in them there won't be any issues at least from my end.
I wonder if any of those spikes were me
There is at the very least 7 day caching so...
eh honestly I don't have issue with the embeds using the API
like none at all
the API should really never be down that often
Just need to add smarter batching and lazy loaikg
this has just been a magical christmas week
pretty sure you've seen all the lights flickering on your end :P
hoping that they don't start blinking red over night
@queicherius seems someone wants a gw2e mobile app https://www.reddit.com/r/Guildwars2/comments/60s2gv/does_a_gw2_efficiency_app_exist/ :P
^ to give some insight into how many things were making >60reqs/minute
omg
(the scales are different)
(but it's like 20%)
it's probably commerce sites which I expect to be making ~200 reqs/minute
5.2k limit vs 6.5k total
the one on the left doesn't distinguish between datacenter
80% hit the limit then
maybe I should change it
nah you gotta add both lines on the right together
that big spike deeefinitely wasn't me
oh wait
yeah
@Eearslya lmao
This ruby library is fast af though
is it using libcurl
5.2k limit vs (6.5k+3.1k) total
yes
if it is, is it still ruby
I mean
maybe
sorta
if a butterfly lands on the nose of a man with a chainsaw
like 55% hit the limit then
and a tree is cut down
at least in that peak
did the butterfly cut down the tree
yes
lol
WELL I'M GOING HOME
AND I'M TAKING MY BUTTERFLIES WITH ME
:wave:
(I'll be back on with more graphs once I VPN back in)
have a nice evening
Really, though, 300 requests to
/v2/items before I could type the next command
you've earned it :)
@lye you should totally publish these graphs somewhere
we'd all love it
api.guildwars2.com/graphs let's go
should we tell lye the new mini isn't on the API yet? :p
whitelisted the item for the mini
should take about 5 minutes to show up
<3
Got 55847 items
Performed ItemCrawlerJob from Async(default) in 37316.83ms
not doing this a long time ago was stupid and i regret it
2500 overflows, oh lord
i wonder if our rate limiting bit keeps track of which IPs are the most offending
Looks like someone's "update-the-whole-api" routine
you can burst 600 requests
like ... maybe it takes more than that?
does the API have more than 125k or whatever objects/
:salt:
correction: someone's poorly-written update-the-whole-api routine
lol.
api genuinely seems faster to respond.. or my housemates are asleep
but then again, that does 9 requests anyway
oh wow, that graph
it lives \o/ Thanks lye :heart:
To retrieve the entirety of
/v2/items you need 280 requests
so you can do that about twice per minute
you'll never need to do that though
nah
i mean, it's cached for 5 minutes server-sided
you literally need to retrieve that once every patch or so
a full update? yeah
you do have to check for new whitelisted items though, but that's just requesting all the ids and diffing it
max 2 requests, if there are new items; otherwise it's just 1 request
Took me 12 minutes just to commit all the items to pgsql once I'd fetched them
then something seems wrong :P
don't think it should take that long
Not sure ActiveRecord likes going that fast
yeah, the only thing I do with items is take the diff and then check the new stuff. Only other option is to query for a specific id to update. I have yet to crawl the whole /v2/items since I first got the whole database
well if a description gets updated
or an icon gets fixed
you wont have it
yup
German translations of names change every patch it feels like
the same happens in Wildstar but the translations are (mostly) better than in GW2. some of the bad translations however have a slight GW2 feel, makes me wonder if they have the same translators
There were 39269 name changes since I started tracking
wow
@Archomeda if you wanted, you could always add info on what happens when you hit rate limit for your reddit post
eh... i did mention the 429
fuck I'm blind
<.<
:smile:
i made a silly thing, but i'm not sure if it works as intended bc i'm dumb. (basically a self-refilling token bucket)
https://gist.github.com/codemasher/10ee438b0dc7a3da9bc8d83040829a60
https://gist.github.com/codemasher/10ee438b0dc7a3da9bc8d83040829a60
thanks Obama Gitter
the 39269 item name changes were only german btw, this is the full table:
| Language | Renamed Items |
|---|---|
| de | 39269 |
| en | 256 |
| es | 12738 |
| fr | 9474 |
you keep a lot of interesting metrics
:chart_with_upwards_trend:
@Eearslya are you doing it all within one transaction?
@lye pretty sure ActiveRecord is doing a SELECT and INSERT per item
if it's all within one transaction it might cut down on lock contention
(rather than one transaction per select/insert)
how do you put 57k items in 1 transaction
BEGIN;
SELECT * FROM items WHERE id = 93;
INSERT INTO items VALUES(...);
SELECT * FROM items WHERE id = 94;
INSERT INTO items VALUES(...);
COMMIT;
You can use
ActiveRecord::Base#transaction which accepts a block to execute for the transaction
so
def scrape()
Item.transaction do
actually_scrape()
end
end
Oooh.
with
ActiveRecord::Base#transaction it's automatically COMMIT'd at the end of the block if no exception is thrown; if an exception is thrown the transaction is ROLLBACKd.
Anyway having it all in a single transaction lets you do stuff like
def scrape()
Item.transaction do
delete_all_items()
refetch_all_items()
end
endprovided that refetch_all_items throws if an unrecoverable error occurs during the fetch, that will atomically update all item records in your db with the latest build
(and since
ActiveRecord::Base#save uses a transaction internally -- it means that you'll have one big transaction rather than 70k little ones)
Why is someone fetching
/v2/currencies?lang=ko
lol
...
well...
why is
ko listed as a valid language
Possible locale 'langId' values include:
en
es
de
fr
ko
zhko and zh are practically useless in our use cases :P
ko just returns ErrUnknown
and pretty sure that
ko doesn't even return valid stuff
since it's not actually a valid language
zh actually returns data
well, it's not always chinese is it?
I thought
zh was using the chinese translations for things
that's what
zh is yeah, but pretty sure i came across things in the past that just showed the english translation
some items only exist on one side or the other
@lye I'm curious, do you actually use Rails, or did you just google all of that on the spot
so the NA/EU-only items probably just aren't translated
@Eearslya ruby is gross
I've been liking it
hmm, fair lye
oh wow, I just used a discord bot that does "paging" result
paging result?
So I play VainGlory on mobile, and this bot can give Match data right, and shows a single match at a time
you lost me
they use the reaction arrow
to navigate through the matches
oh wow, thats actually quite clever
oh
lol what a hack
so when you click on a reaction, the bot updates the text
that's actually amazing
interesting XD
I know right!
dam son
@Nabrok time to up your game!
lol
That's interesting, but limited, since one person can only add a specific reaction once.
Can the bot remove a reaction?
unless removing it counts as a same action
that's true too
bot should be able to remove a reaction yes.
well the match opens with already 1 vote in right and left arrow
so if you click it goes to 2
removing it should also trigger an event
and click back it goes to 1
@Darqam it's the same action if you click it back(remove)
so it flicks between 1 and 2
so it could still work
(or more if more people press it)
I see. there's a lot of things that can be done using reactions like that.
So yeah, thought i'd share that here, since a lot of you are working on different bots :)
something to keep in mind, yeah
From the top of my head : polling (use reaction 1, 2, 3... to select an answer) and confirming participation (for raids, for example)
@darthmaim do you have it easily accessible if/when a en api entry gets updated? I'll eventually get my own version of it running (maybe), but until then if you have that, I'd love to be able to regularly check if anything changes as it happens
no sadly not
I'm only tracking item renames
hm, is even just that something I could occasionally check on?
if not, no worries
I could probably add some sort of API for that
honestly, if it's even just a webpage I need to load myself I'd be fine with that. It's something I could easily manually check once a week or something. I doubt EN stuff changes that often
gitter why
got the link for now, I'll take a peak at that
that are all item name changes since sometime in 2013
alright, thanks. Just want to make sure things are still ok for names at least on the en wiki (they really should be fine I think)
oh wait I think I already have an API
because I was generating those lists for the german wiki
lol, that csv file has 4 messed up entries
all renamed items since some hardcoded date as wiki table
if you sort it by name you get 4 items that... got renamed to blank?
oh that's pretty
there are a few items that have a blank name
so who knows...
¯\_(ツ)_/¯
That's clever.
Uhhm, that link obviously only has german items
I can add en for you
Not sure what you'd show with that for GW2 though.
Also, how'd they get the text in columns?
nah, the en stuff is on the csv file, so I can just auto-format it in notepad
I'll just steal the idea from your link
@Nabrok https://api.guildwars2.com/v2/pvp/games maybe for this?
@Darqam already done: http://gw2wbot.darthmaim.de/renamedItems.php?lang=en
@Nabrok not sure how they got the text in column
all en name changes since 2015-10-20 22:30:00
lol, that works. faster than I could edit
thanks
Yeah, you could have it go through your pvp game history.
that's.... a lot less than I expected (en name changes that is)
yeah
you sure that's everything? the csv has 256 en entries, that doesn't look like 256 lines on that link
On my old guild website I used to gather everybody's pvp data and look for matching game ids to display anything that was a guild match (i.e. 2+ in the same match).
all en name changes since 2015-10-20 22:30:00
the others where before that
oh I see, my bad
changed it to include all
but the german generated table will crash the wiki with that many entries
@Nabrok those bot responses are richembeds
well, looks like there is at least one that wasn't caught by the wiki (chicken feathers)
thats why I always hardcoded the date of when I last copied it to the wiki
gotcha
currently i have this
just an example
@Archomeda nice, looks like that's a fairly recent feature (november 2016)
eh, possibly?
it's definitely some time after the webhooks were introduced
yeah, wasn't around in V8, so it's a V10 or V11 thing
hammerandchisel/discord-api-docs#167
hmm so what are the commands/styling to get that?
it's API only
depends on the library you're using for a bot
Like webhooks, bots and user accounts will be able to use rich embeds.
isn't that just users?
it's "bots and user accounts"
user accounts can use them as well, but you have to run a bot behind it
Actually it's "webhooks, bots and user accounts"
lol
ok
you have "bot accounts" and "user accounts", there's no limitation :P
but you can only do this by using the api directly
i still want a quoting ability...
but yeah, e.g. discord.js has this: https://discord.js.org/#/docs/main/master/class/RichEmbed
kk
archomeda, you can have quoting with selfbotes quite easily ;)
yes true
cba to set it up though :P
and referring to a specific message is still impossible
I guess it's probably time I updated the discord.js version.
even though messages have a snowflake, you can't link them :(
which is odd given the search links them
yeah...
It just annoys me that it requires node 6, is all.
node 6 is lts now
migrate :P
I know, but it's still not in default ubuntu etc
use nvm
Yeah
i dislike using the default repos actually for this
the release process in both ubuntu and debian just takes too long
and debian release process is just horribly outdated imo
but that's just my 2 cents
@Darqam discord uses special formatting for linking channels, users and roles:
- channels: <#...>
- users: <@...>
- roles: <@&...>
tbf, there's room for linking messages really, like: <!...>
yeah...
probably a planned addition that hasn't hit yet
yeah, just like the channel categories we've been waiting on for 3 months now :(
except that one has been announced
purge channel of messages is also awaited by a few I think
those two i'm waiting for them
categories more than mass deletes
https://api.guildwars2.com/v2/items/48762 lol why. Everything else related to wvw season 1 is called
WvW Season One ....
the bot I have right now let's you archive the pass X messages and outputs you a json. but I'd like the same for delete
i also want the ability to modify the first text channel... you can't delete it or change the "read messages" permission because it's the default text channel >_>
I know a few bots that offer purging on a few criteria
time, user, media
the issue is, atm a bot can only delete 100 messages at a time, and if you start spamming deletes for a whole channel, it's api abuse and bot might get banhammered
yeah, it's a long process i suppose
also, hammerandchisel/discord-api-docs#208
oh right, that too
yeah, there was that issue in the past where it caused too much fragmentation when people were mass deleting messages :P
I do love their blog post
Anyone happen to know if the ids are the same between
/v2/commerce/transaction/history and /v2/commerce/transaction/current?
Checking on my local server now; I can't make heads or tails of T-SQL
no they are not lolol.
rip
man I really want to write gw2api.com
maybe I can take a few days off work or something smh
doing order reconciliation would be trivial with postgresql
maybe I should just secretly install postgresql on the api cluster
and write libpq bindings for our javascript contexts
:P
i'm like
only half joking though
there's a lot of low-hanging fruit; e.g. getting per-build per-resource histories as jsondiffs
if you make it open source, we can all write it
yeah and people could submit pull requests and stuff
it'd basically be a pile of go+postgresql
@lye quick question, last changelog you marked /v2/characters/:id/dungeons and /v2/guild/:id/storage as
impl. (what does that mean?) and the api says API not active. That supposed to happen?
is http 206 partial content a normal response header from the api or have I messed up?
thats normal
and did the api always tell you which ids got no response from a bulk expanded query?
huh, where?
oh, now I see
I've never really had a look at the headers response before, and I noticed 'Access-Control-Expose-Headers', which returned "X-Result-Total,X-Result-Count,Warning"
and within warning there's a list of ids which failed
never seen that before
I knew about 206, but have never seen the
Warning header
essentially I've written some dumb code which attempts to get tp prices for some soulbound stuff
@queicherius think he meant 'implemented'; i.e., he's got a format ready, but maybe not the backend support yet?
I thought that was
requires backend support (as the other ones in the changelog)
@queicherius ah by "impl" I meant "requires backend support"
the frontend implementation is live but the backend goes through a much longer release process
@ChieftainAlex 206 means that some ids provided (via the ?ids parameter) were invalid; there's a
Warning header containing the list of ids that were invalid
I don't think
?page can return 206, but might be wrong
@queicherius you sure you're not getting 429's anywhere? I'm not seeing any perfs for gw2e requests and am a bit concerned
oh derp I'm looking at DFW, all your requests are to FRA. derp derp.
Ah okay, thought the
impl might be different
Looks pretty normal on my end :) http://i.imgur.com/ngSZ9AW.png
damn that's fancy af
I think my metrics might be broken
i'm not sure how gw2e is above the FRA line
o.o
why is FRA handling so much more requests than DFW?
:shrug:
I think there are more apps in the EU
oh I got FRA/DFW mixed up
@queicherius also NA/EU WvW isn't high right now
That's 3k reqs/minute right there
:D
okay so what are you using to make that really pretty UI
Is that a lib or is it entirely bespoke
its a bunch of npm dependencies (sparkly + cli-table, mainly) and some sketchy job statistic logic with redis hash fields and lists
but it works better than the job statistics that were natively in the package I used, which tended to lock up redis completely with 1 key per completed job /shrug
oh man that's faaancy
I really have to add logging for all my API requests
@lye Do you track only the last 2 hours or is there something like a 48 hour chart too? Am curious to see if I can spread that load better, or if it is as even as that graph suggests
gimme a sec
FWIW I'm actually not that concerned about it, since it's not like you're making needless requests
once I get a combined
/v2/account I imagine that'll go down by quite a lot
That spike when I turned it back on XD
4AM?
those times are PST I think
Gonna ask for that graph again in 2 days or so :>
I should make
/v2/graphs
Yeah, 11 AM GMT
actually, 12 AM
/v2/graphs please!
Pls do
also, that steady increase on frankfurt doesn't seem to be gw2e only
No, that's the US waking up and realizing sites work again
@Eearslya except the US would be hitting the DFW DC, not FRA :P
Maybe the US all VPN'd to Europe
so... coders in europe going to sleep and starting their automatic api crawling then :p
Or US using EU servers
@lye Well, wrapping the whole thing in a
Item#transaction brought it from 12 minutes to..8. So that helps.
yay
removing the extra transactions from
ActiveRecord::Base#save would probably buy you another 4 minutes
11 items marked as invalid..because of blank names. Dangit.
I don't recall if there's a way to do that though
Yeah, the problem is I have it set to
validate_uniqueness_of :id, so every time it inserts, it does a SELECT first
No way to disable the behavior of
save! apparently: http://api.rubyonrails.org/classes/ActiveRecord/Transactions/ClassMethods.html
I would love to see the client requests of gw2e
queicherius @queicherius may track that with google analytics
Can you adjust that graph to show data points based on user-agent?
you're gonna have to, only sane way for me to get the metrics is with oauth2 :cry:
@Eearslya U-A isn't tracked on my end
Curses
And you can't set ajax user agents
Oh, really? Whoops.
And other headers will trigger a preflight request
Maybe tracking Origin would work
Or a custom query parameter
/v2/items/12542?origin=gw2e
Yeah
eh tracking it that way is kind of flaky
Like the ?wiki=1 the wiki adds
I don't really mind not having per-client request metrics
they'd be interesting to look at but not really actionable :/
looking at options for api tracking since GA limit seems a little low... some start with plans of 750€/month - seems reasonable
wait GA costs money?
I don't ever use GA so ... I thought they were paid with user data :/
it costs money for > 10m data points / month, if I understand their terms correctly
hahaha what is one to do with such puny numbers
ohhh, those graphs
gw2e using 5k req/min wow
the important part is that there's good reason for it
:angel:
is it? ;P
:angel:
elasticsearch all the things ;D
sorry to revisit the above, is generating 206's problematic?
no
k good I can forget it
a 206 is just a partial result https://github.com/codemasher/gw2-database/blob/master/src/Updaters/Items/ItemTempUpdater.php#L68
e.g. the last page of a table