VitalitVaults is healthcare app which takes vitals of patients time to time and shows to doctors which then can diagonize constantly in one screen and predict verdict for patients Since we are trying to archive high availability in our app therfore we have implemented docker and kubernetes so that our app is always available.
Problem Statement: In the era of microservices world, Every single request will go through umpteen numbers of systems in distributed environment for fulfillment. Predicting the cascading application failure in this environment is not easy and we don’t have industry standard solution to pro-actively find/predict the same readily available in open source world.
Need solution for following
- Find cascading application failure ( any new service added in flow should automatically gets registered without any on-boarding or meta data config) Example if F is down, D & E won’t be fully operational in nature
- With telemetry data of compute, Can we find/predict optimal usage of compute and save the cost Example – Based on number of tx processed and CPU / RAM on given node, can we find the optimal usage of node and decommission the node which are either idle (or) processing the less request
- Identify the services which is often not performing as expected
- Visual representation would be an added advantage for viewing #1, #2 and #3