Kubernetes: The Cost of Doing It Wrong

I bike a lot. I bike long distances for exercise, I bike over mountains for meditative calm, I bike with friends to share an activity and a laugh (especially when that friend falls in an injury-free-and-hilarious way). I also bike to the store, to the market, to the bakery, and other places that require me to take a workload home. I’ve strapped a mattress to my back and ridden home with the mattress straight up like a sail and against the wind (not recommended). And I’ve bike packed where all my food, water, tent, and living bits needed to fit on my bike for the entire adventure.

Due to the cargo I carry and the regularity of carrying the cargo, a good bike bag is essential. I’ve had racks in the past, baskets, even a three-wheeled bike for large loads. When I moved to Europe a few months ago, I whittled down my bike collection to just one. The geometry on this bike I took with me (I’m a large person, thus the bike frame is large) makes it such that the bike rack sits almost right underneath me instead of back behind my seat. So when I hang bags on the back rack (called panniers), I kick the bags with every pedal stroke.

Now as I mentioned, I bike a lot. So, you can imagine this drives me nuts. And because I need to carry cargo a lot, I need functional bags. And I keep doing it wrong. I keep changing the rack (frame-mounted or seat post-mounted) and the bags (strap down, clip on, wrap up, velcro down). I think maybe this weekend I finally found a combination that works. But this is after literally six months of trial and error. I feel like an idiot because it’s cost me a small fortune for what should be an easy problem to solve (the store won’t take back my beat up bags). 

Here’s where I say… you’re probably fiddling with your Kubernetes infrastructure too much. Like my bike bags, you’re spending much too much time, and much too much money trying to get it right. Chances are, there was an expert to ask, or a guide, or SOMETHING, that could have kept you from having to learn every lesson the hard way. 

There’s no excuse to have your bike lock bounce out on a rocky path because you’ve resorted to bungee cords to hold things together. There’s no need for that car full of young cute people to laugh at you as you nearly fly over your handlebars when the bag gets sucked into your spokes going full speed through that traffic circle. There’s no excuse for that downtime, there’s no need to have overspent, over-permissioned, crashed, under-provisioned, failed to scale, and that entire department of devs doesn’t need to be laughing at you because “the old way worked just fine before you replatformed.”

Recently, a former employee told me that Fairwinds Insights enables standardization in Kubernetes and that standardization is the sexiest thing you can have in infrastructure. I’m not sure sexy and cloud infrastructure have ever been used together. I’m not sure it works at all. But another friend said something akin to “Kubernetes your team can actually run” — and that has a heck of a ring to it. 

What he means is, a bike with a rack and bags that just works would have saved me an awful lot of time and energy (and I don’t want to talk about the money—the money causes me shame). A Kubernetes infrastructure that works saves you an awful lot of money. And Insights helps you get it right. Sane defaults out of the box gets you going, and the ability to write custom policy helps you extend where you need. A fleet installer means you can put this everywhere you need it. And it’s read only. And it covers security, reliability, and cost. It’s somehow the right panniers the first time. And the right basket. And the right frame bag. And the right saddle bag. You can camp with it, shop with it, and in a pinch, get home with a mattress on it. 

This metaphor between Kubernetes and my bike is now so deeply intertwined, I’m worried IT might get stuck in some spokes, so I’ll stop.

My point is this. There is a cost (not a small cost!) to running Kubernetes wrong. Don’t do it wrong. There are too many good people who have learned the lessons the hard way. No need for you to go do the same yourself. Check out Insights. With our free tier you can kick the tires on a cluster and find out just how far off your setup is, you can see if you’re running known CVEs, check for containers running with permissions they don’t need, and see which workloads are wildly over-provisioned. The cost of making a mistake gets higher the easier it is to remedy—and it’s easy as h*ck right now to get it right, to standardize, to have Kubernetes your team can actually run. 

There’s a cost to doing it wrong, and doing it right is too easy. 

Check out Fairwinds Insights and sign up here.

Use Fairwinds Insights for Free Security, Cost and Developer Enablement In One

*** This is a Security Bloggers Network syndicated blog from Fairwinds | Blog authored by Kendall Miller. Read the original post at: