Bubble down 22nd March - DDoS and other Issues

Yep still down it would appear

when I have the wonderful problem of generating half a million in revenue from my site(s) I’ll gladly pay the $57,000 CAD per year plus a team of people to maintain it.

2 Likes

Would be nice to get some European AWS for latency optimisation but fair play to Bubble for ≈2hrs of downtime in 3 months


1 Like

Some ideas for these scenarios:

  1. Handling scheduled workflow queues is a must
  2. Every Bubble app should have a unique page (similar to the built-in 404), let’s call it “outage”, that is stored off the main Bubble cluster and is displayed when there is a major outage of the main Bubble cluster.

Let’s post solutions rather than flooding the thread with “site is down”… they know it’s down.

16 Likes

3 Likes

This would be nice.

This would be fantastic to have, currently I’m just linking the Bubble Status in all my email reply’s about the site being down.

How would Bubble automatically rout to such a page when down?

Is cloudflare is the problem?

Would most likely require an independent reverse proxy sitting in front of Bubble to make that happen. i.e. Another Cloudflare instance or Azure Front Door that would check Bubble’s availability and redirect to the site down page as needed.

2 Likes

Bubble is back, also I didn’t loose the work I was doing when it crashed

1 Like

mi paginas han vuelto, pero los backend aun no funcionan

I believe this could be a secondary route for the DNS when the primary site fails to load. Every app would have a secondary application on the “outage” cluster that holds a single static error page. If the primary application fails to load, the DNS will route to that secondary cluster’s page based on the same application’s DNS settings.

I’m not an expert in DNS routing, but I believe this is possible.

1 Like

Editor still not working for me

Something like Azure Traffic Manager would work at that level see: How Azure Traffic Manager works | Microsoft Learn DNS might work but the user would hit random timeouts if they they attempt to connect to the IP (primary) that’s not working. Something like Traffic Manager or Front Door would do the backend availability detection and not send anyone to the backend that’s down.

prod is working like normal now for me :+1: :smiley:

1 Like

Yeah things look to be back for frontend and editor :crossed_fingers:

1 Like

It’s not always that easy, and to be honest I have more experience doing that in full code environments where I really can have a lot more flexibility.

One way to implement checks is to identify “error states” that you can look for. If you have a bunch of emails scheduled to be sent, you would want to have a table that logs their successful sending and another task that looks for unsent emails that are past their scheduled date and acts accordingly, etc.

Usually doing this kind of thing requires storing a lot more meta-information about the statuses of the things that you want to happen. It might be easier to answer this on a new thread with some specifics.

Actually, I’ll open a thread for people to share their best practices on this front if I can’t find one already out there.

1 Like

ha vuelto todo hasta ahora, iremos informando

Sheesh! I missed the forum

2 Likes