OnlineOrNot Diaries 13

It's Bastille Day here in France, and before I break out the BBQ and beers (I'm still Aussie, after all), I wanted to reflect on the last two weeks of OnlineOrNot.

If I summarized it in one word it'd be: distracting.

You see, this week was a marketing week, and I had 11 ideas for awesome things to do to keep bringing folks in, following up on the success "Do I need a CDN?" had. Instead, I got baited by the prospect of cheaper compute and being able to build features like ping monitoring, so instead of marketing, I spent the week building OnlineOrNot into a monolith app and using MRSK to deploy to my servers.

I succeeded, by the way, at building the monolith app: it's actually pretty neat to be able to run OnlineOrNot end to end on a single machine (if you remember from a few months ago, OnlineOrNot is a pipeline of AWS Lambda functions stuck together with SQS and SNS). It works great for monitoring a single user's checks, but once you load up 1600 users checks, cracks start to appear.

As a result of running OnlineOrNot in a monolith, uptime checks were randomly timing out, overall latency increased significantly, and made the overall system less reliable. Adding additional monitoring especially for VPSes would cost roughly how much I'd save by moving off AWS Lambda. I spent a week tweaking code, increasing VPS size several times, trying different VPS providers, before eventually giving up, and deciding the AWS Lambda premium is worth paying for the type of business I want to run.

Thankfully, OnlineOrNot runs really well on AWS Lambda, so it's not all doom and gloom.

On a more positive note, this week Abhi wrote in to ask: How do I come up with ideas for what to build?

I started OnlineOrNot as a pretty basic project to meet my needs, all it could do when I launched it was visit a webpage over HTTP, and send an email if OnlineOrNot couldn't reach your page.

My secret for coming up with new ideas to build into OnlineOrNot comes from asking my customers as they sign up: "What's the one thing you were hoping OnlineOrNot could do for you?"

From those basic beginnings, folks have emailed in to ask OnlineOrNot to support their use case, and if it matches my vision for what I want OnlineOrNot to be, I typically build it.