Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Andrey Vasenin
    @avasenin
    in our backend services we are fighting for each microsecond ;)
    Adrian Lewis
    @aidylewis
    The ease of distribution is a really big win, which can usually be a little tricky.
    Andrey Vasenin
    @avasenin
    simple_http_worker is just a example. you could easily write your own worker or patch existing one. For example you could use the current simple_http_worker and change measure units and deploy it to your github and pass it to make_install command
    Adrian Lewis
    @aidylewis
    Could you please tell me what additional change I need to make to the simple_http_worker to enable milliseconds apart from changing units => "microseconds" to units => "ms"?
    Vadim Litvinov
    @merlin86
    mzb_metrics:notify({"latency", histogram}, Latency), => mzb_metrics:notify({"latency", histogram}, Latency div 1000),
    Adrian Lewis
    @aidylewis
    Is this correct?
    % simple_http_worker diff

-    units => "microseconds",
+    units => "ms",

-    mzb_metrics:notify({"latency", histogram}, Latency),
+    mzb_metrics:notify({"latency", histogram}, Latency), => mzb_metrics:notify({"latency", histogram}, Latency div 1000),
    Andrey Vasenin
    @avasenin
    no
    you should replace
    mzb_metrics:notify({"latency", histogram}, Latency)
    to
    Adrian Lewis
    @aidylewis
    Yes. sorry
    Andrey Vasenin
    @avasenin
    mzb_metrics:notify({"latency", histogram}, Latency div 1000)
    dii14
    @dii14
    Hi, I was just starting to use your tool. Is it possible to persist the data gathered and the nice graphs?
    Dmitry Ivanov
    @ethercrow
    Hi @dii14
    what exactly do you mean by persisting?
    dii14
    @dii14
    I am running mzbench from docker and I would like to know where the metrics are stored so that I can have them when I close the container
    Dmitry Ivanov
    @ethercrow
    data is stored in plain files, so unless the whole machine with mzbench server is destroyed or these files are specifically deleted, they are persisted
    dii14
    @dii14
    oh nice, I guess it's just my fault I couldn't locate the files
    is it documented where they are?
    Dmitry Ivanov
    @ethercrow
    metric data is also available through API
    on single bench basis though, not all at once
    dii14
    @dii14
    I see but where are the files actually?I can't really find them
    Dmitry Ivanov
    @ethercrow
    https://machinezone.github.io/mzbench/deployment/#bench_data_dir
    dii14
    @dii14
    oh great thanks, I missed that
    Dmitry Ivanov
    @ethercrow
    welcome
    dii14
    @dii14
    also may I ask if the docker image is working for you?
    it might be because of my proxy firewall
    but puling ridrisov/mzbench takes forever
    Dmitry Ivanov
    @ethercrow
    better ask @parsifal-47
    dii14
    @dii14
    ok tahnks
    Dmitry Ivanov
    @ethercrow
    I haven't used docker image
    Renat Idrisov
    @parsifal-47
    @dii14 I tried this image, was working for me :) do you still have this issue?
    according to dockerhub statistics it was pulled 9 times, so it should be possible
    feel free to create issue anyway
    dii14
    @dii14
    is the ridrisov/mzbench image supposed to support git?
    Renat Idrisov
    @parsifal-47
    not necessary, since it uses rpm with pre-compiled workers, the most probably not, you may find Dockerfile for this image here: https://github.com/machinezone/mzbench/blob/master/docker/Dockerfile it has git if centos7 has it
    dii14
    @dii14
    Hi
    I can see that while running a benchmark in the scenario tab there is this functionality which seems like it allows to send some command to some percentage of the worker nodes
    this seems really cool to me, but what kind of commands can I send?
    Timofey Barmin
    @timofey-barmin
    Hi, dii14
    you can call worker functions there
    dii14
    @dii14
    I tried calling wait and it failed
    Renat Idrisov
    @parsifal-47
    @dii14 what message did you get? "command execution failed"?
    dii14
    @dii14
    On screen message was "command failed"
    in the system logs I get the error:07:02:10.352 [error] <0.224.0> Command run failed with reason {badmatch,{[{'mzb_director16_0@127.0.0.1',[]},{'mzb_worker16_1@127.0.0.1',[ok,ok]}],[]}}
    Command:<<"wait()">>
    Stacktrace:[{mzb_director,handle_call,3,[{file,"/home/ridrisov/mzbench/node/_build/default/deps/mzbench/src/mzb_director.erl"},{line,110}]},{gen_server,try_handle_call,4,[{file,"gen_server.erl"},{line,629}]},{gen_server,handle_msg,5,[{file,"gen_server.erl"},{line,661}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,240}]}]
    My wait was empty but that's not the point, I get the same for a correct command:07:03:45.749 [error] <0.224.0> Command run failed with reason {badmatch,{[{'mzb_director16_0@127.0.0.1',[]},{'mzb_worker16_1@127.0.0.1',[ok,ok]}],[]}}
    Command:<<"wait(3 sec)">>
    Stacktrace:[{mzb_director,handle_call,3,[{file,"/home/ridrisov/mzbench/node/_build/default/deps/mzbench/src/mzb_director.erl"},{line,110}]},{gen_server,try_handle_call,4,[{file,"gen_server.erl"},{line,629}]},{gen_server,handle_msg,5,[{file,"gen_server.erl"},{line,661}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,240}]}]
    Renat Idrisov
    @parsifal-47
    I have reproduced your issue, sorry, the problem was in result code control, fixed now, please update master and try again, thanks for reporting this!
    dii14
    @dii14
    ok cool will try again
    dii14
    @dii14
    Actually I was never able to build from sources behind my proxy
    It is something to do with rebar3 fetching the "registry" package, I suspect that my variables for proxies are not read by rebar3