by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Timothy Perrett
    @timperrett
    on the CP side, you’re alway going to have one registered
    Cory Parent
    @goedelsoup
    that's totally fair, especially given there isn't a good JVM implementation
    Timothy Perrett
    @timperrett
    happy to be wrong by the way, just trying to explore the space and make sure im not missing something
    Cory Parent
    @goedelsoup
    yeah, i'd say all of my questions are of the form, can we make better guarantees on X and is it worth the cost
    Adelbert Chang
    @adelbertc
    ^ this was approximately our discussion yesterday too :laughing:
    Timothy Perrett
    @timperrett
    like all good engineering discussions “It depends”
    im paranoid about operational complexity, as what we have right now “just works”
    so i’d like something that is operationally of equal complexity (within a given approxomation)
    Cory Parent
    @goedelsoup
    for sure. i'm 100% on board and i think you've debunked my devil's advocate questions.
    Adelbert Chang
    @adelbertc
    coolio i am planning on starting work soon then
    Timothy Perrett
    @timperrett
    :fire:
    in parallel our team will be working on the branch deployments RFC
    @adelbertc how are you thinking to refactor? Ideally we would have the agent in another repo, to keep us honest about the boundaries of certain code
    perhaps we should make a repo for it then copy the code over there, then we can keep the 0.14 on the train its on, with another branch (perhaps 1.x) that has these major changes
    i’d like to have a 0.15 with the branch deployment feature
    Adelbert Chang
    @adelbertc

    @timperrett My plan is:

    1. Split the pipeline processor and cleanup pipeline into their distinctive control plane/data plane parts - e.g. "what to deploy" vs. "how to deploy" and "mark as garbage" vs. "sweep garbage"
    2. Come up with the Protobuf data models for the events
    3. Write a reference implementation of a data plane that mimics the status quo
    4. Sink the pipeline processor, cleanup pipeline, sweeper, and deployment monitor into a network port
    5. Migrate the routing cron

    I think #1 and #2 will be good refactorings anyways. I will definitely put the agent in a different repo :+1: Will be sure to give it a cool maritime or astronomy name

    agree on the splitting with 0.14.x and 1.x and stuff
    mhp.
    @miranhpark

    perhaps we should make a repo for it then copy the code over there, then we can keep the 0.14 on the train its on, with another branch (perhaps 1.x) that has these major changes

    just want to +1 that i like this approach too

    Adelbert Chang
    @adelbertc
    Timothy Perrett
    @timperrett
    yeah send your thanks toward @drewgonzales360
    Adelbert Chang
    @adelbertc
    @drewgonzales360 :fire: :fire: :fire:
    Timothy Perrett
    @timperrett
    Oh i’ve had get nelson org added to the Github Actions beta - was thinking i might use actions to trigger the buildkite part to make it auto-release
    then it would just ship itself
    Adelbert Chang
    @adelbertc
    so this fills in that awkward gap we have btwn "i merged to Nelson's branch" and "hey Tim can you make a release pls"
    Timothy Perrett
    @timperrett
    yeahh
    seems sick
    Adelbert Chang
    @adelbertc
    @timperrett did you pre-emptively create this for me https://github.com/getnelson/agent
    Timothy Perrett
    @timperrett
    @adelbertc i did :P
    and a travis-ci job
    Cory Parent
    @goedelsoup
    yooo, so stoked someone finally tackled the TF provider
    Timothy Perrett
    @timperrett
    yeah its going to be really valuable
    Adelbert Chang
    @adelbertc
    i just started learning Terraform too so this is timely :laughing:
    Timothy Perrett
    @timperrett
    we are huge users of terraform - we do everything with it
    Adelbert Chang
    @adelbertc
    we are huge users of terraform but not in any nice consistent way but we're getting there :P
    i linked Atlantis to our terraform guy just now
    Cory Parent
    @goedelsoup
    yeah, all of my infra is in terraform. i'm gonna be very happy to be able to have the nelson admin stuff there too
    @adelbertc you should check out this too https://www.terraform.io/docs/github-actions/index.html it's much simpler than atlantis imo
    Adelbert Chang
    @adelbertc
    @timperrett ^ ? haha
    Timothy Perrett
    @timperrett
    @adelbertc does Github Enterprise have actions? I dont think it does
    Adelbert Chang
    @adelbertc
    i have no clue
    Cory Parent
    @goedelsoup
    from what i understand it's already supported in enterprise
    Timothy Perrett
    @timperrett
    oh really? usually enterprise runs behind github.com, as they bake features there first
    as enterprises are their revenue stream
    Adelbert Chang
    @adelbertc
    lol just got bit with not having blueprints in code migrating a Nelson deployment
    Cory Parent
    @goedelsoup
    that TF provider is gonna be a blessing
    Adelbert Chang
    @adelbertc
    @timperrett are you folks using Prometheus to scrape metrics for Nelson itself? you guys use Wavefront right?
    Cory Parent
    @goedelsoup
    i do myself
    i do prom server per DC + weave/grafana cloud
    Adelbert Chang
    @adelbertc
    i think i am going to have to add some InfluxDB stuff
    Timothy Perrett
    @timperrett
    @adelbertc we use prometheus importer for telegraph, then we send that out to wavefront. We use telegraf for everything