Forum Academy Marketplace Showcase Pricing Features

Latency/Lag Issue - Query Structure, Hardware or Platform?

Hey Bubblers - I hope everyone’s having as much fun as I am in here. Just a quick one for some of the more advanced folks on the thread.

I’m a startup founder who has a working app built in a functional stack (Haskell/Clojure), but I believe visual programming (and Bubble in particular) has real potential to be part of the path forward to an explosion of innovation as more people learn to build. We’re getting to a stage where we’re hoping to put a Bubble-based replacement in front of our existing client (c $50m in annual revenue) as a viable alternative product, but we’ll struggle to get this through if we can’t reduce the lag in page load times. The following link has a video of what I’m referring to: the first load is our existing web app and the second is our Bubble replacement, which looks great but takes quite a while to render.

https://latency.bubbleapps.io/

I’m sure there are a LOT of ways I can improve this app and I’d welcome some guidance from some of the more experienced people here regarding whether I’ve run into (i) the outer limits of my own experience / skill in structuring this, (ii) a temporary issue with system speed due to the impressive user growth the Bubble platform has seen this year, (iii) something else entirely or (iv) a combination of all three.

Don’t get me wrong…still a huge fan of Bubble (after all, where can you build a band new web app in 10 minutes that allows you to sidestep the fact that movies can’t be loaded into this forum), just trying to figure out how we push things forward so people like me who I’d call mid-range in our experience can learn advanced Bubble issues like which structures will render more slowly and which will render more quickly. Maybe an opportunity for some coaching and/or a webinar series on advance Bubble issues? Who knows.

Also, if there are other people who want the full version (including uploader) of the app above to show issues they’re having just respond inline here. I’m happy to build and post a public version (with uploader and Issue page unique to each user) and pay the hosting costs on the videos if it’s helpful to the group.

Best,
Brian

5 Likes

I’ve been looking at this type of issue for my app and the general consensus is to use the Dedicated Plan. Stability and Speed issues are vastly improved. That’s what I recommend.

1 Like

I second @john3.

In the meantime, set a limit of 10-15 rows on any repeating group and then create pagination buttons that quickly loads more data into view.

Another way to improve page load times is reducing number of page loads overall. I accomplish this in all my apps by using a tabbed navigation experience (similar to the Bubble editor) with a page data type set to a database type of thing that holds a majority of the data – you can set conditional statements on view groups that match a Get data from page URL result (ie: …com/page?view=list-of-stuff) so it would be "IF Get view from page URL = list-of-stuff THEN this group is visible? = yes. All of your groups will need to be set to invisible on page load.

With the above approach, my biggest app spends 5-10 seconds loading a sparse view group and begins pulling data from the URL data object (database thing holding majority of data). This sped up every aspect of my apps. Essentially you are reducing calls to the database and eliminating page loads/reloads.

1 Like

Thanks @john3 - super helpful. I’ve been considering the same option. It’s good to hear there’s some kind of consensus out there that it might help to speed things up a bit.

Thanks @philip - I’ve set this page up the same way. It’s multiple “pages” combined in to one that show different information when states are changed. I like the pagination idea - there are definitely areas where I can use it. I’ll start with the Dedicated plan and see how things go from there. Fingers crossed as this is a potentially interesting story of a traditional stack system (100,000+ users, 10,000+ transactions per week) being replaced entirely by a system built on Bubble. Once I get things stabilized and bring the page loads down a bit, I’ll post the full app and story.

2 Likes

I’m also a bit worried about the response times/latency.
Going to the dedicated plan is a big jump, and it would be great to know how big the performace increase will be. Is there any way to do some benchmark tests, and compare dedicated vs shared ?

1 Like

I have the same issues with this as well (my repeating groups with 200 items takes about one minute or two minutes to load) but I know that the Bubble team are working extremely hard to get a big improvement on this performance in the next two months so we´ll see a huge improvement here.

I’ve been using this pingdom tool

But since you’re checking differently configured websites it’s hard to do apples to apples comparison.

I do know that storymodels.co @gregjohnkeegan runs on a shared dedicated which he told me was great. and running the pingdom tool on it gave the following score:

I ran the same tool on my site impactseeker.com which runs on the shared plan and I get a D - but my page size is also bigger.

So maybe run this on your site and see what comes back? And share here?

Plus I asked on dedicated you can configure the following hardware options: Application CPU’s, Database CPU’s, Cache RAM and configure for many more speedy options - of course there is extra pricing involved - but if you care enough about these things then you probably have a service that is valuable enough to pay for it.

2 Likes

My dedicated server’s (storymodels.co) real speed advantage is for CONCURRENT connections due to CPU and memory availability. As we scale, we’ll maintain good speed compared to shared.

The server is EU located FYI.

2 Likes

Nice job man. Great website!!

Hi John, do you have any other example for pages using dedicated servers?

Not really - one is enough for me to know it’s going to be much more stable and faster. Plus you look at the hardware you can configure and it’s all focused on high availability, load balancing, etc…

Worse case you upgrade and try it out and if you don’t like it go back down.

Thanks Dude.

Tested from Stockholm as I’m EU based, performed a bit better.

Put’s me dead centre for average load speeds of all sites. Not bad for Bubble really.

Awesome!

I think getting a B vs every site that they compare is really really good.

You’re talking about sites that are fully configured on non bubble configured hosting solutions.

And I’m sure Bubble will only get better so that A is not too far away I think :slight_smile: