Platform Engineering & IDP Vocabulary

A self-service layer built on top of infrastructure that provides application developers with standardised tools, workflows, and environments — abstracting away infrastructure complexity so developers can deploy, configure, and scale services without needing to be infrastructure experts.

"After building our IDP, application teams go from code commit to production in 15 minutes using a self-service portal — no Jira tickets to the infrastructure team required."

A pre-configured, opinionated, and supported way to build and deploy a service — providing sensible defaults for infrastructure, CI/CD, security, and observability so developers can focus on product code.

"Our golden path gives every new service: a Kubernetes namespace, pre-configured CI pipeline, default monitoring dashboard, and structured logging — without the developer needing to configure any of it."

Synonym for golden path — a well-maintained, recommended approach that teams can follow without needing deep infrastructure knowledge. The alternative is an unpaved road: building custom solutions without platform support.

"Netflix popularised the "paved road" metaphor: teams can go off-road (choose their own solutions) but the platform team only supports and maintains the paved road options."

In Team Topologies, a team aligned to a single flow of work from a business domain — focused on delivering value directly to the customer. The most common team type, empowered to own the full delivery of their service.

"The checkout stream-aligned team owns all aspects of the payment journey: frontend, backend, and database — they deploy independently and are accountable for their service's reliability."

Four engineering performance metrics from the DevOps Research and Assessment: Deployment Frequency, Lead Time for Changes, Mean Time to Recovery (MTTR), and Change Failure Rate — used to measure software delivery performance.

"Our DORA metrics: deploying 12x/day (Elite), lead time of 2 hours (Elite), MTTR of 23 minutes (High), and change failure rate of 4% (High) — platform investments moved us from Low to High performer category in 18 months."

The mental effort required of developers to understand and manage the systems they work with — reducing cognitive load on application teams is a core goal of platform engineering.

"Before the IDP, developers needed deep Kubernetes knowledge to deploy. After: they fill in a YAML template with 5 fields. Cognitive load was the main metric we tracked for platform success."

Operational work that is manual, repetitive, automatable, tactical, and scales linearly with service growth. SRE principle: keep toil below 50% of a team's work to preserve time for engineering improvements.

"Manually rotating TLS certificates across 40 services every 90 days was pure toil — we automated it with cert-manager, saving 16 hours per quarter and reducing the risk of expired certs."

A developer productivity measurement framework covering five dimensions: Satisfaction, Performance, Activity, Communication/collaboration, and Efficiency. Developed at GitHub to capture productivity holistically beyond simple output metrics.

"Using the SPACE framework, we measured developer satisfaction (survey), pull request lead time (performance), and deployment frequency (activity) — the multi-dimensional view revealed that high activity correlated with low satisfaction, pointing to toil overload."