Strangler-fig

Problem

You have a legacy monolith you want to replace, but a big-bang rewrite is too risky.

Context

Rewriting is expensive and error-prone. Any 'V2' project that ships in 2 years fails to keep up with 2 years of legacy changes.

Solution

Introduce a facade (proxy, gateway) in front of the legacy system. Route specific pieces of functionality to a new implementation gradually. Over time, more and more routes go to the new system. Eventually the legacy is 'strangled' and can be retired.

Trade-offs

• Two systems co-exist for a long time — dual maintenance
• Data may need to sync between old and new during transition
• Route configuration becomes a critical failure point

Failure modes

• Migration never finishes — 'temporary' legacy runs for decades
• Data sync fails, systems diverge
• Facade becomes a bottleneck

When to use

• Large legacy replacements
• Zero-downtime migrations
• When you can't take a rewrite outage

When NOT to use

• Small legacy systems — just rewrite
• You have a viable feature freeze on legacy