Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Release 1.22.0 – 15 February 2021 #2985

Closed
timgl opened this issue Jan 18, 2021 · 11 comments
Closed

Release 1.22.0 – 15 February 2021 #2985

timgl opened this issue Jan 18, 2021 · 11 comments
Labels
sprint Sprint planning

Comments

@timgl
Copy link
Collaborator

timgl commented Jan 18, 2021

Context:

  • Automating clickhouse deployments on customers infrastructure + ways of managing
  • Polishing/adding tests

Michael/Marius

James G/Tim

Karl/Eric

  • Sentry bug squashing
  • Issue backlog squashing
  • Slow query reporter
  • e2e test for event ingestion
  • Investigate cypress mock out
  • Typing all the logics
  • Backend test coverage
  • Better demo data

Paolo

@timgl timgl added the sprint Sprint planning label Jan 18, 2021
@Twixes
Copy link
Member

Twixes commented Jan 18, 2021

Here's a list:

  • Be support hero/emir/sheikh/person?
  • Top priority, above all else, is rollout of plugins on Cloud, which we have planned out, but is pending some tests for production confidence
  • Some plugin experience improvements (see Marius' list way below)
  • Settings revamp (this kind of depends on our decision regarding new sidebar by @paolodamico):

There's also implementing the brand new great PostHog homepage (#921), which I'm VERY excited to work on with @corywatilo and @lottiecoxon, this however will be able to begin around Feb 8 at the earliest (pending overall design decisions and then finalized homepage layout).

@Twixes Twixes changed the title release 1.22.0 Release 1.22.0 – 15 February 2021 Jan 18, 2021
@mariusandra
Copy link
Collaborator

mariusandra commented Jan 18, 2021

I'd like to integrate a database explorer in the style of insights into posthog, so we could take the first steps in removing our reliance on metabase and could start cornering another market. Basically #1244

Though practically speaking, I guess there will still be a lot of plugin work to do. Let's see how this sprint goes.

@mariusandra
Copy link
Collaborator

Good Expo support is another often requested feature we should think about: #1174

@mariusandra
Copy link
Collaborator

As alluded to in #3040, it might be interesting to take a few days and clean up some things re: toolbar, including making it slightly less annoying (adding snap to edges for example)

@macobo
Copy link
Contributor

macobo commented Jan 25, 2021

Last release was sort of unstable.

Here's my 2c with some ideas on how to improve:

Improved process for releasing, beta-testers who use a pre-release version.

Performance regression/bug discovery:

  • Trying to "know" which queries are troublesome on self-hosted ahead of time is tricky. The second-best thing is to instrument and ping back home whenever an endpoint takes >50 seconds with the query being run.
    • Ideally with other instrumentation info, e.g. SQL queries + timing of them?
    • Q: Is it okay to do this on self-hosted users who have not opted out w/ the full URL?

Testing:

  • We don't have a true e2e test that checks whether events from $lib actually reach us and are queriable.
    • Initial proposal: Use bookmarklet as part of E2E tests (together with full web server including workers) to trigger events on e.g. google.com, queries the results in the webpage and check whether actions trigger and session + recordings show up.
  • Frontend testing. I think unit tests won't yield too much value w/ kea but integration tests with stubbed API responses might. https://www.npmjs.com/package/@cypress/react would work well here for doing e.g. scene-level testing.
  • Every time we fix something: Save broken query urls to a file, run them against cloud + a postgres instance as part of a release test suite.
  • Quick production test suite which tries to log in/register an account every now and again - smoke testing whether everything is fine.
  • Automatically catch missing team_id filters #3085 Catching missing team_id filters automatically.

Code quality:

  • Enforce js => ts stricter. I've been an offender of this -> help people if they need help doing this.
  • Work on kea typegen to resolve open issues with it.

@EDsCODE
Copy link
Member

EDsCODE commented Jan 26, 2021

General would be good to tackle:

  • better frontend testing along the lines of @macobo's idea. Should be a strong testing process like we have for backend where we patch errors and write tests to prevent regressions
  • staging/prod environments as part of the setup process (less urgent but could be a nice to have for developer experience)
  • slackbot from previous cycle

Specifically for me:

@paolodamico
Copy link
Contributor

Here are my general thoughts for next release:

As for specifically what I'd like to work on next release, I'd like to take it a bit lighter from major features/improvements and focus on:

@macobo
Copy link
Contributor

macobo commented Jan 28, 2021

@mariusandra
Copy link
Collaborator

mariusandra commented Jan 28, 2021

I was quite eager in my previous comments when suggesting:

While I would be happy to dedicate time to them, plugins (aka The PostHog Application Platform) should still be the bigger priority:

You can see the full status on the plugins project here: https://github.com/orgs/PostHog/projects/4

@macobo
Copy link
Contributor

macobo commented Jan 29, 2021

My proposal for todays planning:

Pair up me and Eric. Focus:

  • Testing improvements + Bug tracking/fixing + stability improvements

My additions:

From Erics list:

@Twixes
Copy link
Member

Twixes commented Mar 12, 2021

Released.

@Twixes Twixes closed this as completed Mar 12, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
sprint Sprint planning
Projects
None yet
Development

No branches or pull requests

6 participants