Platform Evolution

A platform-wide analysis of every repository we deploy: what each one is, how healthy it is, how it is coupled and where it runs. Built to support modernisation, prioritisation and roadmap planning for tech and product teams. 213 production systems assessed, traced from Azure DevOps pipelines, and every number here opens the repos behind it.

Estate health, month by month

average score (0 to 100)green reposred repos

Average health score across 213 repos, one point per assessment run. See all change over time.

Where the estate runs

On-prem (145)AWS + on-prem (15)AWS (38)No signal (15)

Derived from pipeline evidence: deployment groups and IIS mean our machines, container builds to Amazon ECS mean AWS. "AWS + on-prem" ships artefacts to both; "no signal" is mostly libraries. Every repo page states its hosting with evidence. How it works.

ZDD readiness: who can leave the three-week release

No excuse57 Ready with guardrails42 Needs work17 A long road38 No chance (frozen core)3 Not applicable56

Zero-downtime deployment readiness, classified from the assessment against the ZDD engineering standard. Click a bar to see the repos behind it.

Shrinking the big release

ZDD is a delivered capability, proven in production: deploy during business hours without taking the platform down. The three-week whole-platform release shrinks by moving volume out of it, starting with the "no excuse" repos, which are held on the train by process, not technology.

"Ready with guardrails" needs named checks first (tests, serialized data, schema care). "A long road" is mostly hosted WCF: strict serialization contracts that cannot run two versions at once. "No chance" is the deliberately frozen legacy core: protect the core, release the rest.

Every repo page carries a ZDD tag and an adoption playbook with its dependencies and gotchas. Read the CI/CD and ZDD whitepaper and the engineering standard in the analysis section.