One deployable, or forty?
Monolith vs microservices is an organisation question disguised as an architecture question.
-
Senior engineers
No juniors. No middlemen.
-
Fixed price
Overruns are on us.
-
Live in weeks
Weekly demos on staging.
-
You own the code
100% of the IP, day one.
A monolith is one application deployed as a unit; microservices split the system into independently deployed services that talk over the network. Start with a well-structured modular monolith unless you have several teams that need to deploy independently or parts of the system with very different scaling needs. Microservices add network failures, distributed data and operational cost, so they pay off only at a size most products have not reached.
Monolith vs microservices, side by side
| Factor | Modular monolith | Microservices |
|---|---|---|
| Deployment | One build, one deploy | Each service deploys on its own |
| Team fit | One to a few teams | Many teams owning separate services |
| Data | One database, transactions are simple | Database per service, eventual consistency |
| Failure modes | In-process calls, fewer moving parts | Network timeouts, retries, partial outages |
| Scaling | Scale the whole app | Scale hot services independently |
| Infrastructure cost | Low | Higher: orchestration, observability, more environments |
| Debugging | One log, one stack trace | Distributed tracing required |
| Best for | Most products, MVPs to mid-size platforms | Large organisations, very uneven load |
Three architectures, three team sizes
-
Modular monolith
Clear modules, one deploy.
-
Monolith + a few services
Split out what hurts.
-
Microservices
Many teams, many deploys.
When to split a service out
-
Very different load
Image processing or AI inference that needs its own scaling and hardware.
-
A separate team
A team that must release on its own schedule without coordinating with everyone.
-
Isolation
Payments or health data that need a stricter security boundary.
-
Different technology
A Python AI service next to a Java or Node.js core.
What we usually recommend
Start with a modular monolith: one deployable, but code split into modules with clear boundaries and no shared tables between them. It ships as fast as a plain monolith and can be split later along the module lines.
When one module's load, team or security needs diverge, extract it into a service. That is evolution with evidence, not a rewrite on a guess. If your system has already split into a distributed mess, see software project rescue.
Signs your microservices are a distributed monolith
- Every feature needs changes in several services at once
- Services share one database
- Deploys have to happen in a specific order
- One service down takes the whole product down
- Nobody can run the system locally
Not sure what you have?
A senior engineer reviews your architecture and code and tells you what to keep, merge or split. Start with a free code audit.
Questions buyers ask us
Are microservices better than a monolith?
Not by default. Microservices solve organisational and scaling problems at the cost of complexity. For most products a modular monolith is faster to build, cheaper to run and easier to debug.
What is a modular monolith?
A single deployable application whose code is divided into modules with clear boundaries and their own data, so it can be split into services later if needed.
When should a startup use microservices?
Almost never at the start. Begin with a modular monolith and extract services when a real need appears: separate teams, very different load or strict isolation.
Can you migrate a monolith to microservices?
Yes, gradually: extract one module at a time behind an API, move its data, and route traffic to the new service. Big-bang rewrites usually fail.
Keep reading
-
Custom software development
Internal tools, client portals, CRM and ERP integrations, and workflow automation. Built by senior engineers on a fixed written quote, owned by you.
-
Cloud migration
Move from on-prem servers, a VPS or a legacy platform to AWS or Google Cloud. Containerised, defined in Terraform, rehearsed, and cut over with a rollback plan.
-
Free code audit
A free, no-obligation codebase audit by a senior engineer. You get a written report: what is healthy, what is risky, and whether to keep, refactor or rebuild.
-
Backend developers
Senior backend engineers in Node.js, Python and Java Spring Boot. APIs, data models, integrations, queues and cloud deploys built to survive real traffic. Monthly or fixed-price.
-
In-house vs outsourcing
Hire developers or hire a team that already exists? A clear verdict, a side-by-side comparison, a fully loaded cost example and a decision checklist.
-
Freelancers vs agency
Freelancer, agency or in-house team for your MVP? A three-way comparison of cost, speed and risk, an illustrative cost example, and when each option is the right call.
An architecture your team can run
30 minutes with a senior engineer. Bring your diagram, or the outage report.