Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    René Ribaud
    @uggla
    @bpetit , I have requeted an account on a HP proliant G9 to a friend. So we could check scaphandre on it. I'm waiting his answer and I'll keep you posted about it, hoping he will be able to get access to us.
    bpetit
    @bpetit
    That's great ! I'd love to hear about how it's going :)
    René Ribaud
    @uggla
    @bpetit, there is something wrong with awx testing. https://github.com/hubblo-org/scaphandre/pull/71/checks?check_run_id=1853744871
    I think a value for the token is missing.
    bpetit
    @bpetit
    Yes I'm working on it. I'm looking for a way to allow people not in the organization (not having the secrets that you see missing here) to trigger the job. Without opening the machine to the whole internet.
    Sorry for the inconvenience, I hope to fix that before tonight
    René Ribaud
    @uggla
    no worries.
    bpetit
    @bpetit

    great feedback and ideas @PierreRust : hubblo-org/scaphandre#75

    Thanks ! I'll have a closer look in the following days and add to the discussion.

    Pierre Rust
    @PierreRust
    hello
    actually, I have some other issues , but I'm not sure if it's really only due to the json exporter
    in the sockets dict for example, I only have uncore energy for socket 0 (and no uncore nor dram) and core (no uncore nor dram) for socket 1
    bpetit
    @bpetit
    Okay, thanks for noticing. I'll have some work on going on the warp10 exporter and then on AMD. I'll try to have a look on the JSON exporter after that.
    @uggla I've asked for a review from you on hubblo-org/scaphandre#76
    No urgency on this, I still have some work to do before merging anyway, but I'd like at least one person looking at my work before I merge stuff, so if you have time at some point to have a look and tell me if you see nonsense it would be great :)
    (I've also moved the get_scaphandre_version function from the riemann exporter to the utils module)
    René Ribaud
    @uggla
    @bpetit , I will have a quick look probably tomorrow.
    René Ribaud
    @uggla
    @bpetit , review and comments done on hubblo-org/scaphandre#76
    René Ribaud
    @uggla
    As a resume, it sounds good, there are minor stuffs to correct and possible improvements.
    I have made a suggestion to try to factorize now. That will be great but it could be done afterward if you wish. Although waiting will make it harder imho.
    bpetit
    @bpetit
    Thanks for the review @uggla . I still have to fix a bug on labels I attach to the data sent to warp10 and then we could discuss about a meeting for pair programming.
    René Ribaud
    @uggla
    Sounds good to me. Any preference for the tool to share code ?
    2 replies
    René Ribaud
    @uggla
    Hello @bpetit , did you had a chance to look at hubblo-org/scaphandre#79
    23 replies
    bpetit
    @bpetit

    Hi !

    This week I'll work on those topics, by order of priority:

    1. generate a clean dashboard to integrate with @PierreRust 's hubblo-org/scaphandre#61 and move forward on the open hubblo-org/scaphandre#53 regarding local docker-compose stack
    2. try to figure out the wrong data in 2-socket CPUs hubblo-org/scaphandre#85. @PierreRust would you like to try a peer programming session on this one ? Or do you prefer to try on your own and have asynchronous discussions ?
    3. work on having proper labels for hubblo-org/scaphandre#84 (docker, kubernetes, ...) to make dashboarding containerized applications easier.
    4. peer-programming with @uggla on hubblo-org/scaphandre#79 to get a clean skeleton for current and new exporters (if both our schedules allow it)
    5. Hopefully release 0.3
    6. I've noticed a tricky and discrete bug in the warp10 exporter I merged last week, I'll try to fix it

    In addition to those topics I'd like to know your opinion on some things:

    Would you be interested in interactive weekly meetings ? I'd share some metrics, previous week snippets, current week topics and of course discuss about what you want in a defined timing.

    Would you be interested in interactive coding sessions, on a given feature or bug ?

    What do you think about moving this chat on a discord server ? This would make peer -programming/screen sharing/interactive sessions easier as there is all the features needed.

    Those are just thoughts right now but I'd like to know what you think about it. I'm interested in making first contributions and bug solving easier (in terms of access to the information like "where should I look if I want to fix this or enhance that ?")

    basic_acid
    @basic_acid:matrix.org
    [m]

    Hello, I'm considering using Scaphandre on a personal project, do you plan adding a way to return more than the first five processes ?

    In fact I was able to get more processes by editing the "Topology" structure and the stdout exporter, but the best would be to add this option to the stdout script (not sure if possible regarding the architecture).

    2 replies
    basic_acid
    @basic_acid:matrix.org
    [m]
    Thank you, I will have a look :)
    René Ribaud
    @uggla
    @basic_acid:matrix.org , you can open an issue. Then I could do it. However if you can do a PR it is even better.
    1 reply
    Tobias Augspurger
    @Ly0n
    Hey scaphande. We are planning our next step to combine the continuous reforestation project with your project.
    https://github.com/protontypes/continuous-reforestation
    At the moment we are writing a blog post on the project launch and the next steps we want to go in combination with scaphandre: https://hackmd.io/Fq-CO7VYRH2-UmcZeazTqw
    Tobias Augspurger
    @Ly0n
    It would be nice to get your opinion / thoughts on the draft. Our goal is to show how open "life cycle assessment tools" in combination with continuous reforestation can be used to compensate emissions within the continuous development process.
    bpetit
    @bpetit

    :clap: This is great ! Thanks for sharing. Don't hesitate to ping some of us here to get some help when needed. We can reschedule our workshop too at some point.

    I'll read that and send you my feedback later today.

    Tobias Augspurger
    @Ly0n
    Thanks, that would be nice. A technical workshop with you guys on the next steps would be amazing. First we need to publish the blog post and do some marketing. I think we can finish that this week.
    @bpetit Don't hesitate to edit the blog post by markdown comments. I made a safety backup just now.
    We will also add you as another author if you like to.
    1 reply
    René Ribaud
    @uggla
    @Ly0n :thumbsup:
    bpetit
    @bpetit

    @PierreRust I'm investigating on hubblo-org/scaphandre#85

    Do you have access to a 2-socket server so that we could test ?

    Pierre Rust
    @PierreRust
    Hello, yes, I have a two socket workstation available. I wanted to investigate the issue on it for some time but I did not find the time to do it yet :(
    bpetit
    @bpetit
    Hi, I completely understand. I'll try to extract some precise pointers on how to fix that and I'll let you try on this machine and give feedback or jump on it and add what's needed. is this ok for you ?
    Pierre Rust
    @PierreRust
    That would be great !
    and btw, regarding your proposition of pair programing on that, I really like the idea :)
    bpetit
    @bpetit

    Great ! We can schedule that starting next week if you want.

    I see there is some work here that seems accurate : hubblo-org/scaphandre#87

    It works perfectly on my setup and seems to replace the function I had suspicions about. Could you try that PR on your machine to see if the bug is still there ?

    bpetit
    @bpetit
    did some of you already play with gRPC and protobuf in rust ?
    René Ribaud
    @uggla
    Just a little as rust-riemann-client is using it.
    It could be a good example to look at it as the "stubs" for riemann are "simple".
    bpetit
    @bpetit
    Great, thanks for the pointer, I'll have a look at it
    :)
    René Ribaud
    @uggla
    The protobuf crate looks pretty good.
    But I would say it is interesting if you have "perf" constraints. Otherwise http is easier.
    bpetit
    @bpetit
    This is not really a choice, I need to discuss with containerd. It seems the most clean way is to do it through grpc/protobuf and the admin socket.
    René Ribaud
    @uggla
    ok in this case, yes you don't have the choice.
    bpetit
    @bpetit

    Hi !

    It seems the linting rules in Clippy have changed a bit in the new version: https://github.com/hubblo-org/scaphandre/runs/2199772368

    It asks us to change struct names like JSONExporter to JsonExporter, CPUStat to CpuStat, and so on.

    I'll proceed on the change on the whole codebase, so be aware of that in your PRs to avoid conflict. The full list of struct names to be changed are in the link just before.

    bpetit
    @bpetit
    Actually I wont, as it may break the API/lib for people using scaph as a crate (even if they should'nt be many). Plus it seems the community will revert that lint rule as many other people are not happy about it: rust-lang/rust-clippy#6974