JupyterHub and BinderHub Monthly Report#

Goal: Continuous improvement

Meeting Date: 21 Mar 2018, 6pm CET#

or your timezone

Welcome to the Team Meeting#

Hello!

If you are joining the team video meeting, sign in below so we know who was here. Roll call:

  • your name / affiliation / github

  • Tim / Wild Tree Tech / @betatim

  • Carol / Project Jupyter (Cal Poly) / @willingc

  • Min / Simula / @minrk

  • Jessica / Jupyter / @jzf2101

  • Erik / IT-Gymnasiet Uppsala / @consideRatio

  • Chris H / UC Berkeley / @choldgraf

Actions#

  1. Help or discussion needed; Agenda Items for Monthly Meeting

    • Open PRs

    • Open issues

      • repo2docker next release 0.6 mid-April

      • z2jh - support for k8 deployments - issue

        • NOTES: two biggest differentiators deploying k8s on baremetal vs. cloud provider is “LoadBalancer” and “PersistentVolume”. Would be valuable to make it clearer which kubernetes objects/concepts we’re using, and which can be “assumed” vs. needed.

        • TODO: create issue to discuss which parts of the jhub deployment depend on which components of kubernetes - create documentation on this.

        • TODO: reach out to the Orsay people about what they’ve done to set this up and what problems they found

        • TODO: reach out to CERN about how installing z2jh went (Tim)

        • k8 incubator: kubespray

    • Kubeflow (JupyterHub and Machine Learning/Tensorflow) Link to JHub issues

      • involvement

      • 15 minute presentation on Friday, April 6 9am Pacific

      • influence direction

      • kaml-d

        • like kubeflow but from RedHat?

      • service bundles and svcat (service catalog)

    • Jessica asks: How do we get better information about interactive computing from Binder?

      • who uses mybinder?

      • what kind of repos are there?

      • Jessica is looking for information to share with the community on how everyone uses binder combining repo info from GH

      • GA gives us an idea of what repos people click and when but not how long they stay

      • can we extract something about “minutes in a binder” from stackdriver logs?

      • we can extend the pages in a user container to include GA in each container

      • investigate GitHub activity stream on bigquery

        • PyCaribbean talk on BigQuery, link?

      • Chris would like help with extracting data out of GA

        • there is a repo: jupyterhub/binder-data

        • Jessica will follow up

      • https://twitter.com/betatim/status/974710693902635010 scientist using mybinder for publication

    • dumping data task:

      • turn this into an issue on jupyterhub/binder, there might already be an issue

      • where to put data from prometheus?

      • how long does the prometheus data last?

      • how do we extract information and what information should we extract?

      • Note that right now prometheus does not log which repos get launched, so we need to figure out the right way to do this, I think Yuvi had thoughts on a path forward there

    • Is anybody interested in a reproducible document collab with eLife / substance.io ? There’s a similar interest w/ a team at the Turing institute in London

      • eLife and substance.io working together on reproducible document format and UI

      • eLife might host their own binderhub

      • maybe interested in helping create tech for publishers

      • Turing institute interested in finding published papers and enhancing their public version

      • Jessica, Tim, Min and Erik would be interested in hearing more

      • Erik figures that that it would be an awesome resource helping out communicating the benefits of mybinder.org to universities.

  2. Decisions needed

    • do we want to commit to a BinderHub release schedule for 0.1?

      • Potentially treat “releases” more as a marker of “which versions work with other versions of other tools” as opposed to defining it by major features. Cut a 0.1 release around the next release of JupyterHub and make it clear that BinderHub is still in big-time alpha version.

  3. Next actions (team)

    • Working toward JupyterHub 0.9 release

      • rough timeline: ~weeks

      • there will be a (short) beta release

News, Information, and Thanks#

  1. Team Metrics

  2. Team News and Informational items

    • Organization highlights

    • BinderHub projects

    • JupyterHub projects

    • Related projects (repo2docker, nbgrader, others)

    • Google Cloud Podcast Interview Monday with Yuvi, Chris, and Jessica

    • Meetings with Orsay in Paris

      • Orsay and OpenDreamKit talking about JupyterHub in education with Min and Chris

    • JupyterDays talk

    • Oxford / Cambridge interest in JupyterHub

    • eLife/substance.io collaboration on reproducible documents?

    • Binder / Turing institute collaboration on “reproducible stars”

    • Had a nice meeting with CZI people at Berkeley, seems like there’s opportunity for collaboration there

  3. Thanks, Things to Celebrate, and anything else

    • Two big milestones that went well: JupyterLab related traffic spike and try.jupyter.org now uses mybinder.org!