When should a UK FinTech move from modular monolith to microservices?

Move from modular monolith to microservices when: (1) multiple engineering teams need independent deployment of different FCA IBS domains (payment team and account team deploying independently — modular monolith deployment blocks both), (2) a specific module has significantly higher scaling requirements than others (payment processing needs 10× the resources of customer support — independent scaling saves cost), (3) FCA PS21/3 Impact Tolerance evidence is clearer with fully independent failure domains (modular monolith in one deployment — a single buggy deploy could affect all IBSs simultaneously). Do NOT move to microservices: because a technical architect read a blog post about Netflix. Netflix has 2,000+ engineers. Most UK FinTechs have 15. ClickMasters recommendation: grow into microservices, do not start with them.

Does NHS DTAC prefer monolith or microservices?

NHS DTAC has no architectural preference — DTAC assesses the product as a whole regardless of architecture. DTAC practical considerations: (1) microservices produce more complex evidence packs (one data flow diagram per service, separate penetration test scope, per-service ROPA items), (2) modular monolith produces simpler evidence (one data flow, one penetration test scope, one ROPA), (3) IEC 62304 traceability is cleaner in a modular monolith (one software system, clear module = software item decomposition), (4) GDS assessment: GDS explicitly warns against microservices complexity for small government digital teams. ClickMasters delivers DTAC-compliant architectures in both patterns — modular monolith produces less DTAC evidence overhead.

Monolith Vs Microservices Vs Modular Monolith

Monolith vs Microservices vs Modular Monolith — UK Architecture Decision (2026)

Updated: March 20269 min read

🇬🇧 UK💷 GBP📊 Comparison🔒 UK GDPR⚖️ UK Law

Direct Answer

The Modular Monolith is ClickMasters' default for most UK regulated projects: single deployment unit (operational simplicity), clear module boundaries (extract microservices later when justified), and lowest operational overhead for teams under 20 engineers. Microservices are right when multiple teams need independent deployment cadence, or when FCA PS21/3 IBS isolation requires independent failure domains. Never build microservices from day 1 on a new product — the domain model is unknown. UK architecture choice must balance operational complexity (regulated sectors require audit trails of production changes), team size, and regulatory isolation requirements.

Factor	Modular Monolith	Microservices	Classic Monolith
FCA PS21/3 IBS isolation	Modules map to IBSs (payment, account, support)	✅ Independent deployment per IBS	IBS boundaries implicit — harder to evidence
NHS DTAC evidence production	Single DTAC evidence pack — simplest	One DTAC pack per service — complex	✅ Single DTAC pack — simplest
Team size fit	✅ 5–25 engineers — optimal	20+ engineers (multiple teams)	✅ < 10 engineers
Operational complexity	Low — one deployment	High — many services, many pipelines	✅ Lowest
Database isolation (UK GDPR)	Module per schema (PostgreSQL schemas)	✅ Database per service (strongest)	Single schema — harder to isolate
Extract to microservices later	✅ Easy — module boundaries already defined	N/A	Hard — no pre-existing boundaries
GDS Service Standard compatibility	✅ GDS prefers simplicity — modular monolith acceptable	GDS caution for small teams	✅ GDS acceptable for small teams
ClickMasters default	✅ All new UK products (NHS, FCA, GDS)	FCA FinTech at scale (3+ product teams)	Legacy — not for new UK builds