is there an easy way to identify which of our plugins use Unpkg ?
2 Likes
Josh,
This is so déjà vu of 2023 and 2024’s post mortems. At what point is Bubble going to stop being dependent on third-party libraries? It seems there’s always an “if only” when it comes to explaining outages.
Do the DB shard crashes explain why the editor is barely usable!? (see below 
for what happens dozens of times a day when trying to do any work in the editor.) It may not be obvious as there’s no response but I’m clicking on the font, style , layout etc with ZERO response..
Let’s ignore how tone-deaf this statement is given the dumpster fire of March 2025 in Bubble (especially the editor). Am I missing features in the dedicated hosting plan that make it suitable for mission-critical use cases? Are there queues or race condition and backend workflow error handling I somehow overlooked?! If so, I spent a lot of time building workarounds that weren’t necessary…
4 Likes
6M Apps → 8 Main Databases. Isn’t it time to stop relying on shared resources? or at least have less apps sharing the same resources. Because @josh you are essentially saying that Bubble can’t handle long queries.
It’s time to make dedicated plans more affordable. Paying 3x the highest monthly WU tier “just to” have a reliable server sounds like a commercial trap.
3 Likes
Every app ran on bubble should have stable hosting, you don’t see wix, squarespace, Shopify, and other web hosting platforms essentially saying “our product is unstable unless you pay us $50k/yr”
Bubble has a duty to keep all sites live and stable if you’re going to force platform locked hosting regardless of plan. (Which was a point bubble made to defend the WU switch, it’ll be more stable for all users by going usage based) now it’s “pay us $50k/yr and you can have stability if you deem your app important enough to have better uptime”
12 Likes
Quick reminder to please keep the conversation focus on the topic at hand. Thanks!
@fede.bubble Not sure why my comment was removed regarding the dedicated plans, it’s very pertinent to what Josh is saying. The whole reason he had to make this post is because bubble apps are experiencing server issues
Our app is mission critical and serves over 100k users per month, we are already on the highest tier plan available outside of dedicated. We have lost customers due to outages but we cannot justify a 9x increase on our monthly bubble server costs just so that we’re not on the shared cluster and I think that is something I’d like @josh and the team to hear.
I am asking for the team to consider a lower barrier to entry dedicated plan that can isolate our app from the main cluster, but at a price point remotely competitive with other PaaS providers that won’t leave myself and other businesses wondering if they should rebuild their apps elsewhere because it is cost prohibitive. The frequency of degraded performance and downtime is unacceptable if there’s no lower tier dedicated plan for us to upgrade to, otherwise thank you Josh for the update.
4 Likes
I kind of feel like @josh made it the topic by using a Platform “Stability” Update as an opportunity to upsell. And a dedicated instance doesn’t provide mission-critical usability (at least out of the box without many workarounds), so that makes it even more on point.
2 Likes
Love that extension (no pun intended)
Do you work at Bubble? My question was a technical question to the CEO of Bubble, not you.
I just stay informed by keeping up with the forum a few times daily and Bubble announcements.
2 Likes
I completely agree, and I definitely empathize with the sentiments around wanting cheaper dedicated hosting. One of my personal goals is to get to a place where people buy Enterprise plans because of the enhanced support, features, and growth, not because of stability: I want our shared hosting to have dedicated-level uptime. I think that’s realistic and doable and we’re investing a ton of resources in making that a reality.
All that said, I feel obliged to point out the option, because today, the reality is it is easier for us to offer higher uptime on high price point dedicated installations. I’d like to make sure users who have the means and need are aware of the option, because it could be valuable for them today – that’s the reason I pointed it out. But yes, this isn’t intended to be our long-term posture, and I view improving stability for all our users as a top priority.
8 Likes
Thanks - so @josh why are you not making dedicated plans less expensive? It just sounds like the easiest and quickest action to take to help users.
@fede.bubble that’s a very frustrating post. Talking about dedicated plans to have a more stable product is clearly align the with post “Platform Stability Update”
Also, per the status page, there has been absolutely NO DOWNTIME for the Bubble Editor in 2025. Is that component of the status page accurate (there are some outages in q3 and q4 of 2024 so its not liked this component is totally unconnected to Bubble’s servers?
If it is accurate and there have been no official outages on the editor, then this is all collapsing faster than even the most pessimistic predictions… and a dedicated instance doesn’t even help as the editor times out for me multiple times every single day on a dedicated instance on my 64GB RAM desktop with 2-3 tabs max open
1 Like
I’m in the last steps of finishing my application.
Therefore I’m testing the main logic of my application thoroughly and I’ve run in to two unexpected behaviors of backend workflow action performance:
- Values aren’t saved to Thing although logs confirm, that e.g. Value ‘X’ was written to Thing’s Field ‘Y’, i.e. confirming that the data was written to Thing.
'List a' merged with 'List b' merged with 'List c'only returns ‘List c’, although List a, b, and c are all Lists of all unique elements. This is not the behavior I expect based on what the documentation of bubble says.
For both of these cases I cannot find any workflow errors in my logs, and these cases are not a one time thing but happen again and again reliably at the same action. I’ve gone over the logic of the Expressions over and over again and am 100% certain that the implemented logic is correct (can provide screenshots If someone isn’t taking what I say at face value)
Perhaps there are further Issues I haven’t found yet.
There seems to be a bigger reliability problem in exactly performing the actions the application / bubble platform is instructed to do – maybe connected to ‘addressing problematic queries’ that Josh points out in his recent announcements?
Any one else having similar problems at the moment? It’s really frustrating to have a program not execute exactly what you tell it to do as this would be the thing you expect from a machine.
I submitted bug reports for all these but haven’t heard any constructive feedback from Bubble yet.
Would be great to hear from your experiences
Perhaps privacy rules are blocking?
2 Likes
Yep, it was privacy rules combined with race conditions due to non sequenced execution of the workflow steps. The user entry into a field of Thing that allows the user to see the relevant field to write to wasn’t yet updated when the writing operation happend.
Now I feel very stupid → developer experience 101
Thank you guys
I would also like to know which of the most popular plugins rely on Unpkg.
But we probably have to go through each of them, check the code and mark it. Maybe we can create a megathread and on the long term be less dependent on them..