Timescale Office Hours: Wednesday, July 19, 2023

Conversation & Community Timescale Office Hours
1

Timescale Office Hours Wednesday: July 19, 2023 09:00 PT / 12:00 ET / 16:00 UTC

office-hours
office-hours2418×938 86.2 KB

We look forward to welcoming you to our open office hours on the THIRD Wednesday of the month at midday Eastern time. See above for the time in your time zone.

These sessions are hosted by our developer advocates and community team and offer the chance to meet with other users, ask questions, and receive inputs from other Timescale users in the field. Here’s the standing agenda

  • Introductions with the Developer Advocates Chris Engelbert and Jônatas Paganini :sparkles:
  • Questions from the community
  • Latest updates from TimescaleDB (with occasional demos and product spotlights)
  • Any other business

The session will be recorded and shared here in this post after the event. If you don’t want to be recorded, you are welcome to join and stay the mic and camera off. You can still participate by joining in the text discussions.

Do you have a question for the team?

Please add your question for the team to address during office hours as a reply to this post. Lots of notice is always very welcome! Browse any existing questions to see if you can bring your experience to share in the discussion.

How to join

Office hours sessions are hosted on Zoom. Here’s the link: Launch Meeting - Zoom

Calendar reminder

To add office hours as a regular event in your calendar, bookmark the Timescale Community Calendar

3

Jônatas informed me that one of the toolkit developers may be able to join next Wednesday (yay!) and asked me to post a couple of questions, so here it goes:

  • Could you present some best practices advice, gotchas or do:s/don’t:s regarding the use of uddsketch() and other percentile approximation functions? My own investigations indicate a few not very intuitive results connected to the input value distribution and the max_error/size parameters.
  • If the metric’s distribution is known ahead of time, using the histogram() function could perhaps be a better choice? What’s the best way to include this function in your continuous aggregates and, possibly, rollups, and subsequently deriving the percentile approximations?

Please reach out on Slack (Daniel Reimhult) if you need more information or clarification.

1 Like