Placeholder for postmortem on editor issues last night

Hi all — as many of you pointed out, the editor was unusable or barely useable for many of you last night, and we under-responded and under-communicated. I’m currently working with the team to get a full timeline on what happened, and will share my findings with you all.

My preliminary understanding is that we shipped a regression, had a fix for it, but were unable to deploy the fix because of an AWS outage going on at the same time that broke our deployment system. I am discussing with the team now why this wasn’t communicated more broadly to the community, and what technical measures could have prevented or ameliorated it.



1 Like

For people that thing this doesnt happen with using code… i just spend 3 days debugging a cloudflare worker issue :confused: and it was caused by cloudflare.

Code and no code will always have bugs. Perfect software is a myth. Bubble is already doing good. With the latest investments in people it will only become better!




And I have a similar tale – One of our devs blew away the production database a week ago… He thought he was dropping the database on staging… nope, it was production. Entire day lost… 3 developers making $75,000/year working together. That ~$1800 cost (including opportunity cost) pays for my Bubble for a while. :wink:


I have similar stories… next.js and other libs not working together anymore out of nowhere… 3 days (with 4 devs) to debug and fix and rewrite stuff. Had to upgrade because of security issues… :confused:

@josh Can we agree that the Bubble team needs to bundle any changes to the editor with the engine updates and ship those at the same time so that a Bubble developer needs to upgrade to the next version before the editor changes get deployed to their editor?


Yeah, we have editor versions for a reason right? These type of “hotfixes” shouldnt be standard. I would love if bubble prioritized stability of the editor over pushing new features.


I do hope you do not reduce this to merely a technical issue to be handled @josh, this is at it’s core an issue of processes such as quality control, proper escalations and having everyone understand that Bubble runs 24/7, globally.


Following up on this here: Incident Postmortem and Updates on Reliability