In early 2018, GraphQL was new and hot tech, and I worked for a startup that kept breaking their GraphQL API. One day a field might be a string, the next day it might be an object.
I got mad, so I built a service that snapshot tested the API response every 60 seconds, and sent a Slack alert whenever it failed. I called it OnlineOrNot.
I spent a year trying to sell that MVP to startups around Sydney, but there were maybe a handful of other companies using GraphQL at that time, and they weren't willing to take a chance on a part-time business solving this problem.
Fast-forward 2 years, and several other failed projects, I decide to rebuild it from scratch, but this time as a general uptime monitoring service.
The URL is https://onlineornot.com, I've been iteratively working on it for around 2 hours per workday since early 2020 (I ruthlessly cut features down into 2 hour blocks, and use feature flags to deploy safely).
These days it's more of a status page, with built-in uptime monitoring (and has integrations to other monitors).
---
Folks tend to ask what makes OnlineOrNot special, trying to figure out what the moat is - it's me.
I've worked on the web for Atlassian and Cloudflare, I've seen what works and doesn't work for self-serve web apps. So OnlineOrNot has:
- a business model that won't suddenly fail (I'm full-time employed so that I can work on OnlineOrNot, I'm not going to shut it down for not making enough money)
- docs written in clear English that load fast, and are up to date
- a modern, responsive web UI with errors that don't make you feel dumb
- uptime monitoring for websites, web apps, and APIs that Just Works
I got mad, so I built a service that snapshot tested the API response every 60 seconds, and sent a Slack alert whenever it failed. I called it OnlineOrNot.
I spent a year trying to sell that MVP to startups around Sydney, but there were maybe a handful of other companies using GraphQL at that time, and they weren't willing to take a chance on a part-time business solving this problem.
Fast-forward 2 years, and several other failed projects, I decide to rebuild it from scratch, but this time as a general uptime monitoring service.
The URL is https://onlineornot.com, I've been iteratively working on it for around 2 hours per workday since early 2020 (I ruthlessly cut features down into 2 hour blocks, and use feature flags to deploy safely).
These days it's more of a status page, with built-in uptime monitoring (and has integrations to other monitors).
---
Folks tend to ask what makes OnlineOrNot special, trying to figure out what the moat is - it's me.
I've worked on the web for Atlassian and Cloudflare, I've seen what works and doesn't work for self-serve web apps. So OnlineOrNot has:
- a business model that won't suddenly fail (I'm full-time employed so that I can work on OnlineOrNot, I'm not going to shut it down for not making enough money)
- docs written in clear English that load fast, and are up to date
- a modern, responsive web UI with errors that don't make you feel dumb
- uptime monitoring for websites, web apps, and APIs that Just Works