DORA Metrics & Developer Experience Platform

The Challenge

The Solution: Multi-Layered Intelligence Platform

I designed and built a comprehensive DevEx platform consisting of three integrated components that transformed engineering operations:

Component 1: DORA Metrics Collector

A Python service that acts as a centralised correlation engine between disconnected systems.

Technical Architecture

Implementation Highlights

• →Retry Logic with Exponential Backoff: Handles API rate limits and transient failures gracefully
• →Git Repository Caching: Clones repos once, pulls updates to minimize Bitbucket load
• →Parallel Processing: Collects metrics for multiple services concurrently (20 services in ~30s)
• →Prometheus Integration: Flask HTTP server exposes /metrics endpoint for Prometheus scraping
• →Health Checks: Collection success/duration metrics for observability of the collector itself

Code Samples

#!/bin/bash

# Never fail the pipeline
set +e

# Enable debug if DEBUG=true
if [ "${DEBUG:-false}" = "true" ]; then
  set -x
  echo "🔍 DEBUG MODE ENABLED"
fi

# Check if webhooks are configured
if [ -z "$TEAMS_WEBHOOK_URL_DEFAULT" ]; then
  echo "❌ ERROR: TEAMS_WEBHOOK_DEFAULT not configured"
  echo "Pipeline notifications are disabled until webhooks are configured"
  echo "Continuing pipeline without notifications..."
  exit 0
fi

# Get current QA deployment SHA from ArgoCD kustomization
get_current_qa_deployment() {
    # Clone or update argocd-apps repo (shallow clone for speed)
    if [[ ! -d "argocd-apps" ]]; then
        git clone --depth 1 \
          "https://x-token-auth:${ARGOCD_APPS_ACCESS_TOKEN}@bitbucket.org/org/argocd-apps.git" \
          argocd-apps 2>/dev/null || {
            log_error "Failed to clone argocd-apps repository"
            return 1
        }
    fi

    # Find the kustomization file for this repo's QA environment
    local kustomization_file="argocd-apps/applications/${BITBUCKET_REPO_SLUG}/overlays/qa/kustomization.yaml"

    # Extract the current image tag from kustomization
    local current_tag=$(grep -A5 "name: ${BITBUCKET_REPO_SLUG}" "$kustomization_file" | \
                        grep 'newTag:' | \
                        sed -E 's/.*newTag: *"?([^"]*)"?/\1/')

    echo "$current_tag"
}

Component 2: Intelligent Pipeline Reporter

A 1000+ line Bash script that transforms pipeline events into rich, actionable Teams notifications with smart routing and context.

Core Features

Example Notification

Component 3: Automated Deployment Gates

Intelligent checks that prevent premature deployments and enforce quality standards.

════════════════════════════════════════════════════
❌ QA DEPLOYMENT BLOCKED
════════════════════════════════════════════════════

The following 2 ticket(s) are missing Fix Version/s:

  🎫 PROJ-1001
     https://company.atlassian.net/browse/PROJ-1001

  🎫 PROJ-1002
     https://company.atlassian.net/browse/PROJ-1002

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📋 ACTION REQUIRED:
1. Open each ticket above in Jira
2. Set the 'Fix Version/s' field
3. Re-run this QA deployment pipeline

Business Impact

Technical Highlights

• ✓Built production-grade Python service with retry logic, caching, and parallel processing
• ✓Integrated 4 separate APIs (Bitbucket, Jira, ArgoCD, Prometheus) into unified intelligence layer
• ✓Designed 15+ custom Prometheus metrics enabling comprehensive DORA dashboards
• ✓Implemented sophisticated Bash scripting (1000+ lines) for rich Teams notifications
• ✓Created automated deployment gates with Jira API integration and delta detection
• ✓Deployed as Kubernetes service with Prometheus scraping and Grafana visualisation
• ✓Improved developer experience with actionable notifications and early validation