> ## Content Index
> Fetch the complete content index at: https://www.jannis.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# When Plans Pan Out
- URL: https://www.jannis.io/when-plans-pan-out/
- Published: 2024-06-23T19:27:33.000Z
- Updated: 2024-06-23T19:27:33.000Z
- Author: Jannis Fedoruk-Betschki

Back in March and April, there were some severe outages on Magic Pages. Some led to a downtime of 2-3 hours. That was everything but good.

Maybe you remember, I took some time to explore ways to avoid that in the future:

[Choosing your customers over profitToday I want to talk about the not-so-sexy part of running a business. The stuff that’s often swept under the rug. Because it’s not fun. And it actually eats up your revenue. I want to talk about the cost of scaling. About choosing your customers over profit. When my business,![](https://www.jannis.io/favicon.ico)Jannis Fedoruk-BetschkiJannis Fedoruk-Betschki![](https://www.jannis.io/content/images/2024/04/growth.webp)](https://www.jannis.io/choosing-your-customers-over-profit/)

The solution back then was quite simple: Replication.

Instead of one server doing all the work, every single service should be replicated not once, but twice. So…three servers doing all the work – three times.

Yes, that means, the Magic Pages infrastructure got quite a bit more expensive. But today I got the proof that it was all worth it 🙌

Around an hour ago I got an email from my upstream server provider:

> One of your servers is currently unavailable due to an outage in our data center. We are aware of this and are actively working on a solution.

Well, no solution in sight yet. But since there were still two other servers running the same services, there was just a tiny little downtime of a few minutes. I bet most people didn't even notice.

All that needed to happen was for an IP address to switch over to one of the other servers. Easy-peasy, automated lemon squeezy. I literally just sat there, put a quick update into Magic Pages's status page, waited for a few minutes, and resolved the issue on the status page.

Now, I am still waiting the server provider to solve the underlying issue. But I know: even if a second server goes down, there is still one more holding all the necessary strings.

It's a great feeling when a plan pans out.