Enel X

Due to a lack of technical documentation and some internal departures, Enel X was unable to support and maintain its current software. They approached OBLSK to enhance and optimize the performance of their software in an extremely accelerated timeline.

After developing a deep understanding of Enel X’s tech stack, OBLSK made performance improvements, streamlined internal processes, added new features, optimized code, and cleaned up data models.

By fast-tracking one project from 3 months to 1 month, OBLSK was able to clean up the mess and quickly position the company for growth in new markets.

Challenge

• Enel X couldn’t support and maintain their current software due to a lack of technical documentation and departures of team members
• Lack of new feature development closed off company to new markets and clients
• API required major adjustments
• Performance problems caused slow response times for critical endpoints averaging 10s of millions of calls per month
• Neo4j was not properly clustered causing speed and efficiency issues
• Code architecture negatively impacted the items listed above

Solution

• Quickly developed a deep understanding of the Enel X’s tech stack without internal direction (internal departures left huge knowledge gaps):
  • Investigating and understanding the current codebase
  • Building a local sandbox
  • Deploying to Enel X’s test environments and production
• Made performance improvements while simplifying internal processes to expedite product development cycles
• Made major adjustments to their API
• Fixed Neo4j clustering configuration by separating leader and follower nodes so all instances were utilized in responding to requests
• Explored logs in Splunk to determine calls to the API most utilized, determined slowest response times and max response time outliers, and changed the architecture, code, and database to improve performance
• Cleaned up data models for more clarity, efficiency, and maintenance
• Ongoing Neo4j maintenance 
• Sped up the project timeline by:
  • Working closely with product teams to understand requirements and find answers to questions earlier in the process
  • Planning workloads focused on only the features required for the project and backlogging nonessential work to be revisited after
  • Limiting meetings to only those necessary for making progress against the fast-tracked project milestones
  • Increasing communication across teams related to feature handoff so that teams would start immediately after a feature was ready for them to take control

Result

• Improved performance by lowering critical response times for endpoints that average 10s of millions of calls per month:
  • Average response time decreased from ~70ms to ~4ms for a number of endpoints called most often
  • An average response time outlier decreased from ~750ms to ~30ms
  • A problematic set of endpoints that had been worked on by previous teams without successful improvement decreased from 3-10 seconds to under ~75ms