We Were Just Replacing the Countertops

We were going to replace the kitchen countertops with reclaimed granite. Give some good material another life, freshen up the kitchen, and keep the project manageable.

As a platform engineer, I should know better than to use the word “just.”

We’d already had the counters templated when the outdated dishwasher under the sink became the main driver of the project. It also turned out to be the source of a leak we discovered in the basement, so keeping it wasn’t much of an option.

Neither was replacing it with another one like it. That particular style of compact under-sink dishwasher was no longer available—rightfully so, in my opinion—and finding a sink basin to go over one wasn’t easy either.

So we had a kitchen configuration built around an appliance we couldn’t replace.

I’ve run into that at work. Something old keeps running, everyone gets used to working around it, and eventually it becomes part of the architecture by default. Then you need to change it and find out how much depends on it.

In this case, we also had water in the basement to help move the discussion along.

The dishwasher needed to move. That meant changing the cabinets around the sink. And that led to the sentence responsible for a good chunk of this project:

“While we’re at it, should we replace the rest of the lower cabinets?”

It was a reasonable question. That’s the problem with “while we’re at it.” It usually is.

We were already changing part of the base cabinets. It made sense to look at the whole configuration. And once we were doing that, upgraded appliances became part of the conversation too.

Individually, the decisions made sense. Together, they were an almost full kitchen reno.

I’ve watched platform projects grow exactly this way. You need to upgrade one dependency, which means changing something beside it. Once you’re in there, you see other things worth fixing. Before long, the small upgrade has a much longer task list.

Some of that extra work is worth doing. You still have to decide where to stop.

For us, that was the upper cabinets. They’re still the originals, painted to match the new base cabinets. They work, they look good with the changes, and we didn’t need to replace them just because we were replacing everything underneath.

I’m comfortable making that call with infrastructure. Apparently, I needed to remind myself I could make it with cabinets too.

The countertop templating had to be reworked once we upgraded the dependencies underneath it. It wasn’t a disaster. The plan changed, so the measurements had to change.

Still, I would have preferred to figure out the dishwasher, sink, and cabinet situation before the first round of templating. Same as at work: a little more time checking dependencies up front can save you from doing part of the job twice.

Then came the plumbing, and with it some new skills. An unexpected turn added electrical work too.

I wanted countertops. I got homework.

There’s plenty of satisfaction in learning how to do something yourself. There’s also the time spent figuring it out before you can actually do it. That time counts, even if the task list makes the job look simple.

Platform engineering has plenty of those tasks. “Update the integration” looks tidy until you’re digging through an unfamiliar system trying to understand why it works the way it does. You learn, adjust the plan, and keep going.

The kitchen was no different, although the debugging positions were considerably less comfortable.

At some point, we stopped calling it a countertop project. New lower cabinets, a different configuration, upgraded appliances, plumbing, and electrical work deserved a more honest description.

Keeping the original name wouldn’t make the work smaller. It would just make us wonder why replacing countertops was taking so much effort.

That’s what I’m taking back to platform engineering from this project. Check what depends on what. Be clear about which changes are required and which ones you’re choosing to add. Keep the parts that still work. And when the scope grows, update the plan.

The reclaimed granite was what we wanted people to see. Getting everything underneath it sorted out was what made it usable.

That’s a pretty good description of platform engineering—and apparently how I ended up renovating most of a kitchen.