Best 3rd Party Database

Meta
1

So, Bubble is not going to be adding in features to help us as developers to reduce our WU consumption. Unfortunately, even though when WU pricing was first introduced, Josh stated publicly they had on a list to implement a feature to allow us to reduce our WUs via electing which fields to return on a search, this feature is no longer on their roadmap. Plus, for those of us who have been around long enough, we can recall the only new features around WUs implemented post the pricing change announcement, were implemented within the first month, and nothing else has been done since.

Bubble also still does not provide a public list of all charges, because the one list they publish 12 months ago doesnā€™t include all of the WU charges for all things, failing to meet complete transparency.

Seems like Bubble is content with keeping things as they are without making any improvements aimed at allowing us to build apps optimally around WU consumption. We can see this through Bubblesā€™ complete lack of public guides or tutorials on how to understand WU in regards to the impacts on WU consumption for such things as finding the count of database entries, lack of improvement on the WU overages emails etc. etc.

So, at this stage, for me personally, the 12 months I was wanting to wait to see what Bubble does since the WU pricing announcement has passed, and they have done almost nothing, so the time has come to fully investigate and begin implementing usage of 3rd party database systems.

A lot of others got started on this 12 months ago, so Iā€™m hoping for some advice/suggestions based on others experience with different 3rd party databases, as to which ones are easiest to work with, which have most functions that can help with backend processing to move some backend functions off of Bubble, and which, if any allow us when running API calls to retrieve data to elect which fields to return (because Bubble does charge per character of data returned via 3rd party APIs).

My goals at this point are to implement a 3rd party database across all of my personal apps, client apps, and templates available on the marketplace. It seems essential now to insulate as much as possible against what seems to be a creeping need to potential pivot away from Bubble, and for me, a 3rd party database is the first step toward that.

8 Likes
2

I host bare metal for all my clients databases. I have a cluster of DBā€™s that are all backed up by CRON jobs. Thatā€™s your best bet to save money. Other hand, you have use supabase and what-not, but they are going to get you on your scaling if you donā€™t consider that. This is why I build my databases on their own.

Another simple approach is to host separate DBā€™s for each client, and for those use-cases Iā€™ll also utilize a service called https://neon.tech. This allows me to scale per client as needed.

EDIT: These days I am no longer utilizing Bubble data in anyway shape or form for my own personal use. I do not trust it, nor ever will (least for important data).

3 Likes
3

There are multiple things to consider here like whether you are wanting just a database or a complete backend option as well as how no-code you want it to be. All the viable platforms Iā€™m familiar with will be at least slightly less no-code than Bubble. I donā€™t think thatā€™s a bad thing though as you get a lot more for the extra learning effort.

Let me say this as well, it is definitely difficult to build using a separate backend, especially initially. Bubble makes the build so incredibly easy to do with the expression composer. Iā€™ve not seen anyone build anything that comes close to rivaling that piece of tech.

However, using a separate backend I think forces better building habits and more modular building.

Here are my thoughts on several backend options though Iā€™m sure others can chime it with other options.

Xano (https://www.xano.com/)
This is my personal choice and I use it every day so Iā€™ll have more information about this than the other options.

Xano is a complete no-code backend thatā€™s established, well funded and ships new features consistently. To your question of returning only the data you want, yes, you have full control over everything. Return the data you want, search and sort by connected data, true fuzzy search, etc

You are also in your own dedicated instance on any of the paid plans so you arenā€™t susceptible to issues caused by someone elseā€™s backend.

We have a dedicated cluster with Bubble but Xano runs the majority of our core functionality for literally 15% of the cost of our dedicated. And, if you like, you can self host the enterprise version of Xano for about half the cost of a Bubble dedicated.

Do keep in mind going this route though, itā€™s on you to watch your metrics such as db storage, etc as your instance will go down if you go over.

Supabase (https://supabase.com/)
This is a hugely popular option and there are a lot of people who would choose this over Xano. Iā€™ve used it very little so canā€™t speak to it much. It is a substantially less no-code option as youā€™ll have to code your functions though I think WeWeb is building some functionality in their platform to help build functions in Supabase visually.

Backendless (https://backendless.com/)
Really fantastic platform but definitely low code as you will be essentially coding all your functions using a block coding system. Felt like it at least deserved a mention.

BuildShipApp (https://buildship.com/)
Have seen lots of good things about this platform but havenā€™t tried it myself. They seem to be iterating quickly and have an Ai tool to help build if thatā€™s something you care about. My general take is they are very good for defined projects (with prebuilt connections) but may not be as customizable as some other options.

FastGen (https://fastgen.com/)
A newer platform but one that has a ton of potential. I havenā€™t used it for any legitimate length of time so I donā€™t have a good grasp on all itā€™s capabilities but some things they are doing well is how they retain all the data from the most recent test run of your function to help with building, the ability to set up ā€˜listenerā€™ functions that can be triggered from anywhere, and being able to import any node module directly into the function. I think weā€™ll be seeing them a lot more over the next couple years.

9 Likes
4

I really like backendless, itā€™s like doing a puzzle

5

You have your backend on Xano and still need dedicated on Bubble? Is that just due to uptime?

2 Likes
6

Didnā€™t Josh write a post recently, I think with last monthā€™s update, that the current to-do list is to aim for a super-fast database, which is particularly about selection of returned data? I think that using the bubble database simply makes the development process much faster than using an external database. Then there are the privacy rules, as well as the simple use of the ā€œCurrent Userā€ with caching and co.

I totally agree with you on the WUs though. It is currently a big mystery as to which workflows consume how many WUs. I recently converted an app with a legacy plan. This app has averaged around 30-40,000 WUs per month for the last 6 months. Since switching to a new plan, the app is already at over 80,000 WUs. Definitely without a change in usage behavior. This makes me very thoughtful.

I donā€™t want to deviate too much from your question, but the order of priority for Bubble definitely needs to be as follows:

  1. stability of the platform
  2. transparency in terms of costs, with detailed examination of how and which costs are incurred
  3. better options for structuring database , i.e. queries in conjunction with selecting which fields can be returned. This will certainly automatically result in apps loading faster, the bubble database being less busy and, accordingly, all users being satisfied.

After that, new functions can be introduced.

I hope for improvement and would prefer not to use an external database. Simply because Bubble would then be deprived of its great strength as a full-stack platform.

1 Like
7

We run hybrid. Still access most live data from the Bubble db but need dedicated for data isolation.

8

This is the power of Bubble for sure. I think itā€™s still sufficient for most apps though can be quite pricey.

2 Likes
9

wonā€™t happen soon, will take a couple of years - thatā€™s why weā€™re looking for alternatives.

also no issues with mvp stage apps - but some are processing records in the range of few hundrend thousand to few millions. at the current stage thereā€™s significant performance difference b/w bubble and other tools at even few thousand records

1 Like
10

Iā€™m not sure which post you are referring to, but specifically on the subject of selecting which fields to return via a searchā€¦

@josh publicly stated this was on their list when they first announced the WU pricing change, as that was my very first thought of a feature that NEEDED to be implemented.

image
image1033Ɨ435 55.2 KB

Because of that public statement by one of the heads of Bubble, I felt that there might not be a need to utilize a 3rd party database. In all reality, I decided at that time to wait 12 months and see what Bubble does. Iā€™ve been trying to get Josh to publicly address this feature for the past 3 months to no avail unfortunately. I then reached out to @laura.oppenheimer in a private message 2 weeks ago asking about whether or not this feature was on the roadmap and when it would be implemented. I have not heard back from her, but another user has been able to hear from her.

image
image796Ɨ291 23.8 KB

Since Bubble has removed that feature from their roadmap, it seems that it will never become a feature, despite all of the benefits it would allow for.

As I have some big plans for the next 12 months on Bubble, personally I need to start working on the projects with a clear understanding of what Bubble is or is not going to be doing regarding releasing the feature of electing which data fields to return, as it impacts the way we build our apps considerably. As it has been made clear that they will not release the feature, and likely never will, I now need to investigate 3rd party database options.

Also, with the lack of reliability, poor performance, differences in views between Bubble and the community on the types of features and improvements to be made, consistent flops of new features, for me personally, I need to mitigate the risk of being completely locked into Bubble, and the first step to that is using a 3rd party database.

At the end of the day, Bubblesā€™ editor is great, which they know is their core product. I will be continuing to use Bubblesā€™ editor to build apps quickly, but need to reduce my dependence on Bubble, so it is just natural, to not keep all of my eggs in one basket anymore.

I also have experience with clients I built apps for, leaving Bubble for traditional code the moment they got funding. So in anticipation of a potential future of getting funding and, what seems like an inevitability of the investors to demand rebuilding in traditional code, it seems prudent to be better prepared for that ahead of time and, again, first step for that is an external database.

I am baffled that 12 months after having released this, and only 6 or less months away from forcing everybody onto it, they still have not sat down with engineering to put together a COMPLETE list of ALL things that cost WUs.

One area that came to light recently was that, an unlisted cost of 0.05 WUs for a database trigger scan is incurred EVERY TIME we create or modify a data type entry in the database, whether or not the app has a backend database trigger for that data type.

You would imagine that within the past 12 months, somebody at Bubble would have one, flagged the fact that the price list doesnā€™t include this cost, and two, that it is ridiculous that we are charged for the scan even if we do not have a backend database trigger change for the datatype, and three, that they have not released a feature to either allow us to elect to not have the feature of backend database trigger events or even better, to have a feature for Bubble to know which data types have an associated backend database trigger event and only charge the 0.05 WUs for a database trigger scan if those datatypes with a related database trigger scan are created or modified.

5 Likes
11

Like you I am frustrated with all of the false hope and announcements regarding optimizing WU that came out at the time of the initial announcement and the pretty much silence since. I have some apps running millions of WU and the only interaction I have had with Bubble is with Sales people reach out to me to offer me dedicated.

No assistance, no guides, no help, no understanding on optimization, nothing. Just that they can ā€œhelp meā€ if I get dedicated. I donā€™t think thats quite fair or part of what was said when the new pricing was just announced.

I donā€™t want to be too critical but isolating the users who have managed to scale applications instead of turning them into champions of the brand seems not the smartest move but we are where we are.

The one great thing about the pricing announcement was the amount of time provided before it kicks in. I am thankful for that, me and my team have been going deep into learning Xano amongst other things. Weā€™re excited about the opportunities it provides to use our backend DB for other little niche projects and concepts such as Discord Bots, TG Bots, Flutterflow (Iā€™m thinking the upcoming Bubble mobile release will just be a wrapper and a better mobile native experience can still be designed elsewhere) and its also made us look at WeWeb amongst other tools.

In terms of using an external db, the advantage is many that you can return individual fields which helps with privacy, you can re-use APIs that you create across different projects (tg bot, discord bot, flutterflow etc) so it is generally just a much more robust system. Does it add to development time? Yes. Can you be confident that what you have built is pretty bulletproof with easy error management? Also yes.

It is a shame the direction that is being taken here with the lack of care / communication to those who are still on legacy other than the odd interaction with disingenuous sales people who are ā€˜trying to helpā€™. I will be interested to see what happens once the new pricing kicks in. Weā€™ll actually be able to reduce our Bubble bill quite significantly which I think is fair for making us jump through these hoops.

3 Likes