Let's hustle #2

Sharing a side-hustle idea that you can build a business around

Introduction

Welcome to Let’s hustle. In this new series, we’re gonna talk about development-related side-hustle ideas that you can build a business around.

I put together a small Go program that transcribes the most important technology and business related podcasts and extracts interesting business ideas from them. They discuss a lot.

I’ll also share my own ideas.

Let’s hustle!

SaaSykit

If you want to ship your own idea, today’s sponsor can be very helpful to you.

SaaSykit is a feature-rich Laravel starter kit with everything you need to build an awesome SaaS.

It comes packed with many features like payment integrations (Stripe, Paddle & Lemon Squeezy), beautiful checkout flow, admin panel, user dashboard, auth & social login, ready-made components, stats board, built-in blog, extensive docs and much more. Save 500+ hours building your startup.

SaaSykit is available in two flavours, SaaSykit (user-based starter kit) and SaaSykit Tenancy (multi tenancy-supported starter kit).

Use CSSIMPLIFIED code to get 25% discount. Limited time offer!

I’m going to use SaaSykit for my next idea, because it comes with multi tenancy, and seat-based subscriptions which would be a real pain to develop. It also provides nice looking components which is also a big win for me.

The problem

The other day I was watching Casey Muratori’s famous “Clean” code, horrible performance video. It’s a good one where he goes through the main selling points of clean code and shows how it ruins the performance of an application compared to simple procedural code. It’s a good video. But, of course, he’s a game engine developer. He has a vey different point of view of performance than you and I. Sometimes I envy people like him. So it got me thinking.

The vast majority of performance problems I experienced as a PHP developer working on high-level business applications was related to databases. Or to be more precise, to my poor knowledge of databases, indices, and queries in general.

Almost all applications use a database. And it seems crazy to me how “few” tools we have to analyze and improve the performance.

Just think about the usual methods:

• Running EXPLAIN queries

• Browsing the slow query log

In my experience (which is probably very limited), this is the default “toolset” when it comes to improving the DB’s perfromance. (In fact, I was asked multiple times in job interviews if I know “explain” at all. Which indicates there are developers who don’t know what it is, and don’t use it. Which indicates they probably know nothing about the slow query log. Which indicates they just send a Jira ticket to DevOps to scale up the database when there’s a performance problem.)

And then there are some more “advanced” methods, such as:

• mysqltuner

• sysbench

• Getting lost in performance_schema

• Going crazy over EXPLAIN ANALYZE’s output

• Paying thousands of dollars for APM (application performance monitoring)

Don’t get me wrong. These tools are good, and they can be efficient. But, they are often limited, and hard to understand.

For example, this information_schema query:

select * 
from table_io_waits_summary_by_index_usage
where index_name is not null
order by sum_timer_wait desc;

should show index usages. And it does:

Good luck parsing that information. I’d rather have a slow application.

On the other hand, APM solutions can be quite expensive, and hard-to-setup. For example, Coinbase paid $65M to Datadog. Yes, 65 million. Of course, that’s an extreme case, but cloud solutions can be pretty expensive, generally speaking. (Basecamp famously saves $850,000 per year by leaving the cloud. Which is quite wild. They provide the same service but spend $850,000 less).

So my problem is this: knowing what is slow in your DB, why is it slow, and how to fix it should be easy and cheap to find out. But it’s not.

(I already built the starting point for this idea as an open-source package)

The idea