Estimation template for proxy development

[#1]

Need advice on planning sprints that involve proxy development on Apigee Edge. More specifically, need help on how to approach estimating a proxy for development in one or more sprints. The estimations can vary based on many criteria like developer expertise, complexity of proxy being developed, policies involved, language used, etc.

Are there any estimation templates or best practices that can be shared in this area?

[kkleva]

What I do is try to build out a set of standard user stories that can be used as a guideline for proxy development.

Currently, I have these in categories of API Specification, API Functional Development, etc. Something like this:

API Specification
 - API Specification - SB Target Definition
 - API Specification - Confluence and Markdown
 - API Specification - Transforms and Wiring
 - API Specification - Test Review and Research
 - API Specification - Spec Review
 - API Specification - Test Plan Review
 - API Specification - Test Execution
 - API Specification - Error Scenario Data Setup
 - API Specification - Happy Path Scenario Data Setup
 - API Specification - Blueprint Test Execution
 - API Specification - Create New Jmeter Test
 
API Basic Set Up User Story

Gold Standard Functional User Stories
 - API Caching User Story
 - API Basic CREATE User Story
 - API Basic READ User Story
 - API Basic UPDATE User Story
 - API Basic DELETE User Story
 - API Analytics User Story


Gold Standard Non Functional User Stories
 - API Basic Traffic Management User Story
 - API Northbound Security User Story
 - API Southbound Security User Story


Gold Standard Technical User Stories
 - API Error Handling User Story
 - API Header Management User Story
 - API Logging User Story
 - API Payload Transformation User Story


Gold Standard Node.js User Stories
 - Node.js Module Basic Set Up User Story
 - Node.js Module Specification - Package Dependencies
 - Node.js Module Specification - Reference Documentation
 - Node.js Target Specification - Package Dependencies
 - Node.js Target Specification - Reference Documentation
 - JavaScript Unit Test Setup User Story

For each Gold Standard story, I'll build out the 'what', 'how' and 'definition of done' for each story. From here I'll apply some average story points and hours it typically takes to accomplish each task.

The key here is to start to collect actuals of each task. From here you'll start getting an idea how YOUR SPECIFIC TEAM is able to finish. Over time, you'll start getting the confidence required to provide good estimates.

I've been using this general approach for a few years to create over 70 proxies. It works!

[#3]

@Kristopher Kleva Thanks for sharing, this is useful.