Skip to content

2023-02-23 launch check in

General check in

  • Two working days left!
  • Kriti extended current iteration to 2023-02-27
  • Some items in Launch nice-to-haves are moved to next week’s iteration (e.g. load testing optimizing reflection)
  • Only left things in the current milestone that seem to have a reasonable chance of getting done
  • This view is up to date:
  • What if demo site is not ready for lots of traffic?
    • We still launch. If it crashes, then oh well. Demo is just a nice-to-have.

Users & Permissions

  • QA plan is coming along.
  • Rajat will review the QA plan.
  • Anish and Rajat will handle the QA work.
  • Will be done by tomorrow
  • Mukesh and Anish are doing some review work for users and permissions
  • Reviewers are encouraged to push changes to the PR just so we can get the PRs merged quickly


  • Still some outstanding work for documenting users and permissions
  • Kriti is reviewing users & permissions docs
  • Sean reviewing README PR

Installation & Deployment

  • Deployment type 2 is working, tested locally, tested on GCP test server. PR is submitted, awaiting review.
  • Documented differences (in same PR) so users can understand the permissions needed by the user.
  • We have hired someone from Upwork to write some more installation docs

Live demo

  • We need to discuss some design/UX issues
  • Timing for improvements to live demo: We are okay pushing some imprvements to the live demo after the release, but we want to get as much in before the release as possible
  • Load balancer is in progress
    • Need to figure out how to determine when to scale.
  • We’ll chat about load balancing implementation details later
  • We’re hoping to have some improvements completed tomorrow


  • Waiting to work out installation issues before we integrate the docs


  • Still need to QA the upgrade process


  • no updates


  • no updates


  • Issue with “Schemas (0)” displaying sometimes for demo users and local installations
    • Sean experiencing this locally too
    • We’ve had some reports from live demo users that have hit this bug
    • Sean will hop on a call today with a back-end engineer to troubleshoot

Priorities per person


  • Finish all the QA testing tasks
  • Takeover #2520
  • Community work


  • Recreating schema missing/uncreatable bug
  • Solving schema missing / uncreatable bug
  • Writing up release plan
  • Keeping an eye on Deployment type 2 reviews, working towards merge


  • Test in-app updating (with real Github releases)
  • Fix Arxiv dataset in demo
  • Work on reflection in demo


  • Update style for docs
  • Alternative design for CTA


  • Help move forward installation and deployment
  • Help with upgrades if needed
  • Demo analytics testing
  • Publicity work


  • Review deployment type 2
  • Work on Load balancer for demo server
    • Have a basic loadbalancer setup
    • Figure out the health check to scale theinstances
    • Point the mathesar domain to the loadbalancer IP


  • Test Upgrades
  • On-call for Users & Permissions issues
  • Add more Analytics events


  • Review the QA plan
  • Do the remaining QA items for users and permissions
  • Review related changes on users and permissions documentation.


Upgrade testing procedure

Participants: Pavish, Dom, Sean, Kriti Summary: figure out how to test upgrades and who’s going to work on it.

  • Preparation:

    • Dom: Make sure there are two releases (both on Github and Dockerhub)
      • with different versions strigs/tags in code, in git and in docker
    • Dom: Figure out steps to switch between releases in dev
      • so that we can trigger in-app updating and then “rewind”
  • Testing process:

    1. Dom Call the upgrade endpoint to double-check that it still works
    2. Pavish Install that release locally and test upgrade has been updated accordingly

Live Demo navigation changes UX

Participants: Sean, Kriti, Ghislaine

Sean, Kriti, and Ghislaine looked at in-progress work and finalized UX for changes to the live demo banner. Related issues:


Live Demo load-balancing strategy

Attendees: Kriti, Mukesh, Dom, Brent, Sean

  • We need to figure out when we trigger adding a new server in our autoscaling policy
  • Lots of back and forth discussion
  • Because of reflection bugs, demo server will be unresponsive once we hit a certain number of DBs created, even if there’s only a single user using it.
    • It is unclear if we can use memory / CPU / number of processes as an indicator of this
      • Dom thinks we can’t
      • Unresponsiveness should be able to be correlated with something
    • We could add an endpoint that indicates if there are too many databases and use that as a healthcheck to trigger scaling up
    • It would be best if we didn’t need to add a new endpoint and we could use an existing indicator
  • Conclusion: Mukesh will load test as needed and identify the best metric
    • Dom pushed the latest tests to mathesar-scripts repo
  • Setting up infrastructure is the most important thing, we can change autoscaling policy anytime

Schema bug - debugging session

Participants: Brent, Sean, Kriti, Mukesh

  • Related to
  • Sean screenshared and showed bug happening locally
    • This is because demo mode is broken, created to track it
  • Brent identified issue happening for production, because of collectstatic
  • Ideally, we should be building static files as part of the release, not during install process
  • Brent will figure it out tomorrow, he has enough info