#metabrainz

/

      • Nyanko-sensei joined the channel
      • 2016-11-30 33518, 2016

      • D4RK-PH0ENiX has quit
      • 2016-11-30 33521, 2016

      • Gentlecat
        how did you move it there? can I get a copy?
      • 2016-11-30 33542, 2016

      • Gentlecat
        wanted to see if the upgrade works since I found some new docs
      • 2016-11-30 33540, 2016

      • Nyanko-sensei has quit
      • 2016-11-30 33540, 2016

      • dpmittal
        Gentlecat : ping
      • 2016-11-30 33556, 2016

      • Gentlecat
        hello
      • 2016-11-30 33539, 2016

      • Gentlecat
        I've seen your submission btw
      • 2016-11-30 33546, 2016

      • Gentlecat
        will get to it soon
      • 2016-11-30 33553, 2016

      • dpmittal
        Gentlecat : i was asking was there anything else which you guys told me to add or to rectify and i didnt do it ?
      • 2016-11-30 33519, 2016

      • D4RK-PH0ENiX joined the channel
      • 2016-11-30 33533, 2016

      • chirlu has quit
      • 2016-11-30 33538, 2016

      • chirlu joined the channel
      • 2016-11-30 33510, 2016

      • Leo_Verto joined the channel
      • 2016-11-30 33529, 2016

      • Lord_of_Codes_ joined the channel
      • 2016-11-30 33543, 2016

      • Lord_of_Codes has quit
      • 2016-11-30 33553, 2016

      • Lord_of_Codes_ is now known as Lord_of_Codes
      • 2016-11-30 33548, 2016

      • Gentlecat
        dpmittal: I can comment here if you prefer
      • 2016-11-30 33514, 2016

      • dpmittal
        Gentlecat : yeah sure
      • 2016-11-30 33530, 2016

      • lazka joined the channel
      • 2016-11-30 33505, 2016

      • Gentlecat
        in some places you mention just music reviews, like "how to write a good music review on CritiqueBrainz"
      • 2016-11-30 33544, 2016

      • Gentlecat
        but it's not just music anymore. we have places, events, and more types are likely to be available in the future :)
      • 2016-11-30 33509, 2016

      • Gentlecat
        I'm not sure if I understand this sentence: "Choose a language which expresses your emotions rather than facts."
      • 2016-11-30 33533, 2016

      • dpmittal
        Should i change "music reviews on CB " to just "reviews on CB" wherever needed?
      • 2016-11-30 33545, 2016

      • Gentlecat
        that would be better, yes
      • 2016-11-30 33557, 2016

      • Slurpee joined the channel
      • 2016-11-30 33557, 2016

      • Slurpee has quit
      • 2016-11-30 33557, 2016

      • Slurpee joined the channel
      • 2016-11-30 33517, 2016

      • dpmittal
        And with that sentence i meant that choose a language in which you feel free to express your emotions/thoughts instead of choosing a language in which you can just write factual data
      • 2016-11-30 33516, 2016

      • Gentlecat
        it should be fine to write just facts
      • 2016-11-30 33547, 2016

      • reosarevok
        Well, just facts is hardly a review
      • 2016-11-30 33512, 2016

      • Gentlecat
        right
      • 2016-11-30 33556, 2016

      • dpmittal
        I'll remove that sentence then
      • 2016-11-30 33558, 2016

      • Gentlecat
        maybe suggest to choose language which person finds the easiest to express themselves in
      • 2016-11-30 33518, 2016

      • Gentlecat
        another thing that would be nice to do is to replace he/she with they
      • 2016-11-30 33532, 2016

      • Gentlecat
        I'd also make capitalization consistent
      • 2016-11-30 33556, 2016

      • Gentlecat
        you have titles like "Live performance (event) reviews" and "Summing Up the review as a Whole"
      • 2016-11-30 33539, 2016

      • Gentlecat prefers the former version (just first word)
      • 2016-11-30 33514, 2016

      • Lord_of_Codes has quit
      • 2016-11-30 33531, 2016

      • Lord_of_Codes joined the channel
      • 2016-11-30 33503, 2016

      • Gentlecat
        also, the structure should be cleaned up a bit
      • 2016-11-30 33511, 2016

      • Gentlecat
      • 2016-11-30 33544, 2016

      • demonimin has quit
      • 2016-11-30 33538, 2016

      • zas
        bitmap: îng
      • 2016-11-30 33546, 2016

      • zas
        ping* ;)
      • 2016-11-30 33550, 2016

      • bitmap
        pong
      • 2016-11-30 33559, 2016

      • zas
      • 2016-11-30 33515, 2016

      • zas
        interesting things from those graphs
      • 2016-11-30 33529, 2016

      • zas
        look at what happened at ~14:00 gmt yesterday, a bump in ws usage, so we hit the current global rate limit
      • 2016-11-30 33542, 2016

      • zas
        and check load on servers
      • 2016-11-30 33559, 2016

      • zas
        ws increase had almost no impact on load
      • 2016-11-30 33530, 2016

      • zas
        now select only 200 - (main website)
      • 2016-11-30 33528, 2016

      • bitmap
        yeah
      • 2016-11-30 33537, 2016

      • zas
        from 14 to midnight we have a traffic increase, but look at load on hip/ludwig/trille/bootsy
      • 2016-11-30 33530, 2016

      • zas
        it continuously increased during this period, it makes me thing something is very wrong
      • 2016-11-30 33541, 2016

      • zas
        s/thing/think/
      • 2016-11-30 33533, 2016

      • demonimin joined the channel
      • 2016-11-30 33556, 2016

      • zas
        if we look at numbers ... we have serious load peaks with only ... 35 req/s on 4 servers... that's awfully low
      • 2016-11-30 33519, 2016

      • zas
        in the same time the ws is handling 10x more requests and passing from 200 r/s to 350 r/s had almost no impact on load
      • 2016-11-30 33553, 2016

      • bitmap
        hmm
      • 2016-11-30 33528, 2016

      • bitmap
        it'd be nice if we could log URLs where the response time from the backend exceeds a certain #
      • 2016-11-30 33555, 2016

      • bitmap
        the slower ones are probably generating the most load
      • 2016-11-30 33528, 2016

      • zas
        yes, i added the mean request time to graphs, reload
      • 2016-11-30 33545, 2016

      • zas
        ws is consistently low, even on load
      • 2016-11-30 33525, 2016

      • zas
        can you log slow requests on backend side ?
      • 2016-11-30 33503, 2016

      • bitmap
        I'm not sure how, I was hoping nginx could
      • 2016-11-30 33501, 2016

      • zas
        on gateways ?
      • 2016-11-30 33506, 2016

      • bitmap
        yeah
      • 2016-11-30 33534, 2016

      • bitmap
        or I guess we could add some custom code to mbs in the catalyst request cycle
      • 2016-11-30 33525, 2016

      • zas
        i would prefer a solution on the backend service, rather one on the gateways
      • 2016-11-30 33512, 2016

      • bitmap
        ok
      • 2016-11-30 33529, 2016

      • bitmap
        we can add something there I guess https://github.com/metabrainz/musicbrainz-server/…
      • 2016-11-30 33501, 2016

      • bitmap
        but just log the URL at certain intervals instead of killing it
      • 2016-11-30 33507, 2016

      • zas
        yes, looks simple enough, we can do it on one backend server for now
      • 2016-11-30 33532, 2016

      • zas
        then provide it as an activable feature
      • 2016-11-30 33545, 2016

      • zas
        with configurable threshold
      • 2016-11-30 33514, 2016

      • bitmap
        I'll try adding something on hip
      • 2016-11-30 33559, 2016

      • bitmap
        log if it takes longer than the mean request time? (0.5s)
      • 2016-11-30 33546, 2016

      • bitmap
        ah, alarm only works with whole seconds
      • 2016-11-30 33548, 2016

      • zas
        you could start with 10s and see how many lines you get, then decrease
      • 2016-11-30 33514, 2016

      • bitmap
        alright
      • 2016-11-30 33512, 2016

      • zas
        response time and load are linked, so you'll get more lines during peak hours
      • 2016-11-30 33529, 2016

      • dboys has quit
      • 2016-11-30 33504, 2016

      • regagain joined the channel
      • 2016-11-30 33537, 2016

      • mihaitish has quit
      • 2016-11-30 33508, 2016

      • bitmap
        just a few indexed search pages with 5s
      • 2016-11-30 33526, 2016

      • bitmap
        lowered to 1s...lots of indexed /search pages, and quite a few artist and release pages
      • 2016-11-30 33544, 2016

      • bitmap
        tail -f /root/slow-website.txt
      • 2016-11-30 33548, 2016

      • github joined the channel
      • 2016-11-30 33548, 2016

      • github
        [metabrainz.org] gentlecat opened pull request #241: Update Flask-Admin requirement to 1.4.2 (master...admin-update) https://git.io/v1Oxc
      • 2016-11-30 33548, 2016

      • github has left the channel
      • 2016-11-30 33513, 2016

      • dpmittal
        Gentlecat : by problems did you mean the capitalisation problem or something more than it? I have changed the structure of the wiki as you suggested and tried to make the doc a general one and not restricted only to music
      • 2016-11-30 33550, 2016

      • bitmap
        the indexed searches are a bit of a problem since they have to wait for the search server, but I'm not sure they are actually producing much load while they're waiting
      • 2016-11-30 33545, 2016

      • Gentlecat
        dpmittal: https://chatlogs.metabrainz.org/brainzbot/metabra… (and the next 5 messages)
      • 2016-11-30 33519, 2016

      • ruaok
        I also don't understand why the indexed search is so slow.
      • 2016-11-30 33517, 2016

      • zas
        ruaok: why do you say indexed search is slow ? response time is very low, but when indexes are rotated
      • 2016-11-30 33553, 2016

      • zas
        mean response time of upstream servers for search is < 0.02 s, but rotation causes huge bumps of course
      • 2016-11-30 33532, 2016

      • ruaok
        there are still loads and loads of timed out queries happening. we time out anything that takes more than 1s.
      • 2016-11-30 33558, 2016

      • ruaok
        and the slow queries really pile up during indexe rotation.
      • 2016-11-30 33518, 2016

      • lazka has quit
      • 2016-11-30 33536, 2016

      • zas
        ruaok: what is the status code returned when those are timing out ?
      • 2016-11-30 33501, 2016

      • Mineo joined the channel
      • 2016-11-30 33540, 2016

      • github joined the channel
      • 2016-11-30 33540, 2016

      • github
        [metabrainz.org] gentlecat closed pull request #241: Update Flask-Admin requirement to 1.4.2 (master...admin-update) https://git.io/v1Oxc
      • 2016-11-30 33540, 2016

      • github has left the channel
      • 2016-11-30 33551, 2016

      • zas
        we have 150-200 req/minute on each search server ending with 400 status code, i wonder about those
      • 2016-11-30 33529, 2016

      • ruaok
        in java speak: REQUEST_TIMEOUT_EXCEEDED ("Search request timed out -- please wait a moment and try again")
      • 2016-11-30 33554, 2016

      • ruaok
        the 400s are clients not properly escaping content in their queries, resulting in invalid queries.
      • 2016-11-30 33517, 2016

      • zas
        ruaok: if you provide me a status code i may be able to draw the number of expired queries per minute (if the code isn't used for anything else)
      • 2016-11-30 33529, 2016

      • ruaok
        still looking... java. :(
      • 2016-11-30 33539, 2016

      • ruaok
        SC_REQUEST_TIMEOUT Status code (408)
      • 2016-11-30 33514, 2016

      • ruaok
        so, it should be possible to make that graph and I'd really appreciate that.
      • 2016-11-30 33505, 2016

      • D4RK-PH0ENiX has quit
      • 2016-11-30 33532, 2016

      • D4RK-PH0ENiX joined the channel
      • 2016-11-30 33512, 2016

      • dpmittal
        Gentlecat : i have updated the doc again trying to address all your suggestions :D
      • 2016-11-30 33532, 2016

      • Gentlecat
        👍
      • 2016-11-30 33510, 2016

      • CallerNo6
        dpmittal, I think you picked the hardest task :-)
      • 2016-11-30 33556, 2016

      • github joined the channel
      • 2016-11-30 33556, 2016

      • github
        [metabrainz.org] gentlecat opened pull request #242: Update the README file to match updated Docker setup (master...readme) https://git.io/v13I3
      • 2016-11-30 33556, 2016

      • github has left the channel
      • 2016-11-30 33545, 2016

      • CallerNo6
        dpmittal, Gentlecat, maybe this is something lost in translation, but "Being biased in your review rather than being objectively critical or praising" is a mistake in a review? That seems like exactly what a review is.
      • 2016-11-30 33528, 2016

      • ruaok signs the SiriusXM contract
      • 2016-11-30 33519, 2016

      • bitmap
        CallerNo6: +1
      • 2016-11-30 33514, 2016

      • bitmap
        maybe if the band doesn't show up to your event you can be objectively critical, idk
      • 2016-11-30 33542, 2016

      • Gentlecat
        I'm also not sure about "Writing either entirely positive or super negative reviews."
      • 2016-11-30 33558, 2016

      • Freso
      • 2016-11-30 33507, 2016

      • Gentlecat
        for example, I can't come up with anything negative to say about https://critiquebrainz.org/release-group/4344b426…
      • 2016-11-30 33547, 2016

      • yvanz
        Freso: Not for now, because the URL cleanup component is in the middle of a large refactoring.
      • 2016-11-30 33536, 2016

      • aron_kexp has quit
      • 2016-11-30 33550, 2016

      • aron_kexp joined the channel
      • 2016-11-30 33529, 2016

      • bitmap
        right, telling people how they should feel about something seems wrong
      • 2016-11-30 33518, 2016

      • bitmap
        that seems to be saying to avoid writing reviews about stuff you feel very strongly about, or to temper your views first :P
      • 2016-11-30 33556, 2016

      • ruaok is probably late to the party
      • 2016-11-30 33502, 2016

      • ruaok
        did you folks see this?
      • 2016-11-30 33503, 2016

      • ruaok
      • 2016-11-30 33544, 2016

      • kepstin
        Someone posted that here last week
      • 2016-11-30 33550, 2016

      • kepstin
        I just shared it
      • 2016-11-30 33552, 2016

      • CallerNo6
        fashionably late!
      • 2016-11-30 33515, 2016

      • Freso
        ❤️ CC
      • 2016-11-30 33536, 2016

      • CallerNo6
        stanislas_, ping. Do you have a copy of your paper from https://codein.withgoogle.com/archive/2015/organi… - I'd really like to re-read it.
      • 2016-11-30 33557, 2016

      • dpmittal
        CallerNo6 , Gentlecat : "writing a partial review being biased towards a particular artist of the event/album " is this a mistake w while reviewing?
      • 2016-11-30 33524, 2016

      • Freso
        I would assume a review to be partial.
      • 2016-11-30 33551, 2016

      • ruaok
        that is the whole point of CB.
      • 2016-11-30 33500, 2016

      • ruaok
        if you want impartial, go read wikipedia.
      • 2016-11-30 33505, 2016

      • CallerNo6 doesn't understand how bias is a mistake in a review. Maybe you mean that reviews should critique the music, not the musician?
      • 2016-11-30 33533, 2016

      • CallerNo6
        (which I still wouldn't totally agree with)
      • 2016-11-30 33503, 2016

      • CatQuest
        I think it should be *the writers* real aoppinion. not "the crowds" (even if that opinion isn't the "most accepted" one)
      • 2016-11-30 33531, 2016

      • CatQuest
        (ie justin beiber sucks/is awesome)
      • 2016-11-30 33533, 2016

      • CatQuest
        writing "JB is.. ok" might be considered weird .. idunno
      • 2016-11-30 33519, 2016

      • CallerNo6
        CatQuest, I guess it's arguable, but IMO if you're reviewing an event, you're reviewing more than just the performance.