Image builders, such as Docker, build images automatically by reading the instructions from a file (Dockerfile).

A Dockerfile is a text file that contains the commands that build an image. Crafting Dockerfiles to build efficient images comes with its fair share of problems and pitfalls.

Problems include, but are not limited to:

-Updating external packages,
-Using the "latest" tag
-Running multiple services on a single container
-Running containers as root

The effect is inefficient, mutable and insecure container images that take a long time to build, verify and deploy.

The solution is to adhere to standard best practices when authoring Dockerfiles and building container images. Dockerfile reference templates automate knowledge and simplify the adoption of these practices. This approach has multiple benefits, including:

-Best practices: templates codify and automate the adoption of best practices
-Reusability: a single template can be used across multiple projects
-Collaboration: common templates with well-known structure can be easily maintained by larger number of developers with different skill levels.
-Time-savings: a well-documented template provides the guardrails that speeds up the creation of container images.
-Efficiency: templates based on best practices tend to produce more efficient container images

This epic covers the work to research, develop and publish reference Dockerfile templates that can be re-used across projects and developers.

Templates will be treated as first class software development artifacts and will be subject to the same lifecycle of source code, including source control and code review cycles.