Lifelog in decades

Fuck Heroku.

What was supposedly a 15min database upgrade turned into a 15h downtime, and a whole Sunday’s worth of panic-driven coding.

This is why I hate using managed hosting.

There’s always some upgrade, some patch, some dependency going on. Developers in a large tech company needs work in order to justify their enormous salary, so their work becomes my work too… all in the name of security or performance.

Now, the previous database no longer works with Lifelog, and luckily I was able to spin up a duplicate in my VPS over at Digital Ocean.

For the time being.

As much as I hate to refactor Lifelog—it was what made a bunch of us writers homeless to start with, and what inspired Lifelog to be built—I suspect there’s no escaping it now.

Heroku forced my hand.

Now I have to rebuild Lifelog from ground up, on land I can be more sure will leave me the fk alone.

Maybe it’s a good thing. It needed to be done, eventually.

Lifelog’s infrastructure was built on a tech stack when I was still learning to code. Much compromises and half patches were made to get it out in the wild.

But now that I know how to code.
I can get this built in a way to last decades.

Decades.

Because that’s how long we’re planning to keep writing, right guys?