We are planning to migrate 50+ legacy applications to AWS. What KPIs should we be tracking to ensure the migration is successful and the apps are performing better in the cloud?
3 answers
During a migration, you need to track metrics in three categories: Business, Financial, and Technical. Technically, you must benchmark your on-prem latency and throughput before the move so you have a baseline. Post-migration, look closely at CPU/Memory utilization to see if you are over-provisioned (which wastes money). Also, monitor "Error Rates" and "Request Latency" via an APM tool like New Relic or AWS CloudWatch. If your latency increases post-migration, it usually points to a database configuration issue or a sub-optimal network route between your new cloud VPC and remaining on-prem services.
How do you accurately measure the "Total Cost of Ownership" (TCO) during the migration phase when you are essentially paying for both on-prem and cloud infrastructure?
I recommend tracking "User Experience Scores." Sometimes the metrics look green, but users complain of slowness due to new authentication hops in the cloud.
Great point, Janice. Real User Monitoring (RUM) is vital to ensure the technical success of the migration translates to a better experience for the actual customers.
That "double-bubble" cost period is tricky, Louis. We track "Migration Velocity" alongside our cloud bill. The goal is to shut down on-prem servers as fast as possible. We use tagging in AWS to track exactly which migrated apps are driving costs. This allows us to show leadership that while the cloud bill is rising, the on-prem maintenance and power costs are dropping, eventually leading to a lower overall TCO.