Last week, I wrote about some changes that I was making to Siteimp's monitoring module to better solve a more stringent monitoring problem. It was a monitoring problem that I had recent experience with - I built a small local AI model backed application that I use as a guide to recipes while I cook. When the model goes down, the chat just thinks so I never really know whether I just asked something too complex or whether my model died.

But back in the day, I was heavily involved in surveillance cameras. There was a brief point in the mid 2000s when I was a bit of a name in storage and retrieval. It was a big problem then as bandwidth had not quite caught up to the needs exposed when you had 16 surveillance cameras running in a business and the owners need offsite backup and remote view capability. I made a lot of serious mistakes and killed that business off within a few years - it turns out that distributing to retailers is a heck of a big problem and I just couldn't solve the problems that problem created.

However, I made some good friends and stay in touch with some of them. One friend of mine got into the legal cannabis industry and owns a small production facility. Or, small by the standards set by his ambition - it's a heck of a great operation. His operation uses a mix of cloud and on premises based equipment to automate as much of the farming process as possible - the margins in the legal cannabis industry are so small right now that automation is a big way for legal producers to provide a good product and compete with the black market in terms of price. Back in the day, I built a surveillance system for him and his now husband for a vertical farm they started.

It turns out that monitoring is a source of concern for him. And while Watch wasn't quite ready for mass consumption, I spent a marathon session in front of my machine and got a build out to him. And so, what started off as a quick catch up about family, my kid and our baseball obsession turned into a beta trial.

And so, Siteimp Watch is officially in beta. Through building it out, I had to come up with a few more features that made their way back to Siteimp. Our robots.txt policy is much stronger and we introduced two user-agent strings that will stop any Siteimp scan or monitoring request before they start. It's a great step to building a more honest product that we won't control once it gets released - we run scans and monitoring on a user's own hardware so there is a small potential for mischief.

Since Watch is a very focused product, it gets rid of some of the annoyances that make Siteimp Monitoring great, but not quite great enough for hardcore internal uptime monitoring. I can check stuff that runs in my network once a second and get notifications to (in this case) Discord.

And of course, I have to write a metric crap load of support documents. Metric crap load is a unit of measurement... if you've ever had to write user facing tech support documents for a product, you are familiar with it. There are two crap loads in a metric shit tonne and 20 imperial pounds in a metric crap load. Or something like that. :)

Happy Sunday.

About the Author

Greg Hluska is a performance-obsessed developer focused on improving Core Web Vitals and real-world speed. He built Siteimp to make optimization faster, easier, and more reliable. Learn more about performance on the Performance Optimization blog. Currently working primarily on the Fitness Tracker application and 78solutions, Greg is busy. But not too busy to spend his free time with his child.