Try Our Sweet New Databases


Update September 18, 2017: Everyone has now been transferred to the new databases, and everything appears to be working without a hitch. There is one remaining step, transferring the non-document data (emails, settings, etc), which will happen on Wednesday, Sept 20th at 9:14pm Pacific Time, and will require a few minutes of scheduled downtime (we are expecting 20 minutes of downtime, but it could take as long as an hour). You should still be able to use WorkFlowy in offline mode during the downtime.

Update September 11, 2017: We have now begun the process of transferring everyone to the new databases. At this point, around 300 existing users have transferred to the new databases, and we’ve also been putting new users on the new databases. In sum, that’s around 4,000 people on the new databases with no reports of data issues.

WorkFlowy is close to outgrowing the largest single database on offer by Amazon Web Services, where we host the application. To deal with this we’ve been engaged in a pretty involved project to split all the data we store across many different databases.

Internally we’ve been using the new database cluster for about a week, and we’ll soon start transitioning everyone over. Before we do that, however, we want a few intrepid souls to try the new database setup and report if they’re having any unexpected behaviors when using WorkFlowy. We haven’t had any issues with the new setup yet, so hopefully you won’t either, but we’d still appreciate your help kicking the tires.

Here’s a link where you beta test the new database cluster:

Try WorkFlowy’s New Database Cluster

Please email any issues you find to, thanks for your help!

17 thoughts on “Try Our Sweet New Databases

  1. (I know this may be way premature) – So far so good! within first 15 seconds of upgrade it already appears faster! My largest list expands in 3 seconds instead of 30 seconds.

  2. Load times unchanged for me (about 9 seconds for full reload), but save time seems twice as fast, now closer to half a second.

  3. Probably a silly question, but most of my use is through the Android app. Now I’ve ticked the update box on the web, does the app also switch over?

  4. I have done everything I can think of to break your new database… but all I come up with is: loading time cut in half, lightning-fast filtering/ searching, incredibly smooth transitions between lists while navigating, near-instantaneous list sharing times, seamless loading for many hacks like browser extension tweaks (Stylish, Tampermonkey) including list background colors, iframe, images, etc. etc.

    And I must say that with over 100,000 lists I never really thought that the current database was lagging at all. It’s all good… but when you are introduced to faster than fast, your idea of “fast” is recalibrated. The new database is spectacular.

  5. Like others, have noticed the difference in load speed. No issues. Not a power user, like frank.dg and others, but I certainly use it as my do everything toolkit.

  6. I have been trying to reach out and didn’t find any relevant info on website. Hope you can help me here. Is there a way to show time / chronological dependence between two items without indenting one under the other? Example – I have two tasks/projects – Task A and Task B, at the same bullet list level. Meaning one is not a sub task of another. However Task B can only start after Task A is completed. How can I indicate this chronological dependency in WorkFlowy? Please help.

    1. Hard to do right now, but you could write under Task B, “Waiting on Task A”, and include a link to Task A (which on clicking, you would visit). There should be better ways of doing this in the future.

      Do you have an idea of how you’d want that to work? What’s your dream scenario?

    2. tags or simple [1], [2], [3]… or [A], [B], [C]…at the beginning of each bullet. Tags are nice for dependencies that might not be next to each other and allows for worry-free moving and easy tag explosion, while numbers and letters make for a nice readable outline when an ordered task list is preferred.

      Convenient tag format: `#p-ProjectID-TaskID` as in `#p-53-A` or `#p-TestDependencies-A`. This way one can search for `#p ` to find all projects and for `#p-TestDependencies ` to find all tasks within a project

    3. Omnifocus has ways of dealing with that in the program (e.g., you can set thing up so you don’t even see Task B until Task A has been completed). WF is more freeform. When I came over to WF from OF, this bothered me. But over time, I’ve adjusted. It’s a series of tradeoffs. OF had more of that structure, notifications, etc. But it was also slower. WF is very fast and the search is also very fast. I also like that it’s platform agnostic (although I with the mobile app was a bit better). That’s why I switched. I’m in the Apple ecosystem all the time.

  7. Jesse, above you asked for our ‘dream scenario’ for dependent tasks: perhaps something like GTDNext, where two dependent but equal tasks show as siblings, but the 2nd task doesn’t show until the first task is completed, or deleted. Say, for example, you have a project of home remodeling of 10 sequential tasks. You set them up as 10 children of the project in Workflowy, but show only sequentially. You only see Task 1, then when it is done (or deleted) then Task 2 automatically pops up. So you see Task 2 until it’s completed, and so on …. Remember, you asked for our dream scenario. (Thanks!)
    ps. I’m a first-time poster here, but have used Workflowy for a few years now, and I love it.

  8. Fwiw, you could do worse than looking at Omnifocus for task dependencies. Sounds like it works similarly to GTDNext.

Comments are closed.