By constructing an all-in-one project management system, OpsSpark dedicated team has helped Pulse allow its users to accomplish more with less time and effort.
There were many unused components in the system, and each deployment took up to an hour. There was also no separation between resource allocation and the new code implementation (deployment).
Here are some challenges:
- All provisions and deployments take up to one hour to finish.
- Provision and deployments open failed itself, no way to stop it
- Many development environments are in trouble, the formal DevOps team only cares about production environments
- Many redundant components which are can be merged into one, obsoletes are not clean-up
- Site performances (i.e loading time) are not optimized with caching and geographical distribution
- No separation between resource allocation and deployment
- There are no incremental backups of the database and restore strategy
This is how we approach and solve PulseMS’s problems:
- Separate provision and deployment, make sure deployment is less than 15 minutes
- Identify key factors on AWS cost explorers then reduce
- Review cloud formation stacks to remove redundant obsoletes components and resources
- Build CDN strategy for caching and optimizing static and dynamic contents
- Build backup & restore database strategy
- The system is in good shape
- Cloud costs reduced 25%
- Provision time reduced 20%
- Deployment time reduced to 80%
- DB Backup and Restore are implemented and tested well
- The system becomes more compact, all dev and production environments are taken care of for better project development