Should you start with a monolith and scale later, or go straight into microservices?
Tech giants like Amazon and Netflix swear by microservices, b...
For further actions, you may consider blocking this person and/or reporting abuse
monolith can solve 80% of your problem, the rest is not about the architecture
Most small to medium size business will almost never need a true micro services architecture. There are exceptions maybe. Like if you really going grow to enterprise scale due to a guarantee of contract or corporate buyout with the expectations (and guaranteed funds) to grow the business.
Absolutely! For most small to medium-sized businesses, a well-structured monolith is often the best choice. Microservices introduce complexity that isn’t always justified unless there’s a clear need, like rapid scaling, enterprise-level growth, or specific workload separation.
Your point about guaranteed contracts or corporate buyouts changing the equation is a solid one. I’ll be updating the article soon to highlight this perspective so others can make more informed decisions. Appreciate your input, thanks for sharing!
Correct
Agreed 👍
Nice, interesting topic.
I won't go for microservices too early on. If you don't plan on building a mobile app, then just go with a monolith. If you plan on a mobile app, just split your app into an API and a frontend layer.
The bottleneck is often the DB layer with web applications. When your traffic grows you can just load balance DB servers, cache servers, and even web servers so long as your monolith is stateless (you store images on S3, sessions in Redis, and other user data in the DB).
A microservice only makes sense when you need to squeeze maximum performance or when part of your application requires special hardware. For instance, if you run scrapers, these require running a headless browser.
You don't want to run those processes on the same server as your primary website because they are resource-intensive. Instead, you could build a microservice to use Cloudflare workers. They spawn and self-destruct on the fly depending on what resources you need at any given time. This could be way cheaper than just running a dedicated server as well. Since serverless worker runtimes are memory and disk space-sensitive, you don't want to shove your entire monolith in there, hence why a lightweight microservice makes more sense.
Great insights! You’ve outlined some key considerations that many developers overlook when deciding between a monolith and microservices. Your point about handling high traffic with proper load balancing, caching, and database scaling before jumping into microservices is spot on. Also, the example of offloading resource-intensive tasks like scrapers to Cloudflare Workers is a fantastic real-world application.
I’ll be updating the article soon to integrate these valuable points so everyone can benefit from them. Really appreciate your input, uthanks for sharing your expertise!
Awesome, thank you for the kind words. Glad it was useful.
My pleasure
Most of the time Microservices are overkills.
Yes. Monolith is great for start up, however, we cannot underestimate the role of microservice in big company who need separation of concern in infrastructural management
Nice.
I was unsure which option would suit me better, but after reading this, I am now convinced that a monolith is the way to go.
Thanks for the post!
You are always welcome🙏
Great posts.
Thank you