We recently our migrated our application from VM and in-memory Database to AKS and Azure Postgres flexible server. After the migrations, we feel DB is not stable and when monitoring the CPU utilisation it was going 90% sometimes and Django application is not responding because DB is taking longer time to respond. I know this is not straight forward issue and need more informations to debug. But i need two suggestions here, i'm not expecting the solutions,
How to handle the applications down if DB is not responding, i need a workaround to handle this scenarios.
Our team is optimising the queries, but i'm not feeling that's the solution and our DB size is very small (Less than 1GB) but we have Memory Optimized, E2ds_v4, 2 vCores, 16 GiB RAM, 512 GiB storage. Still it's not able to handle the requests, i'm not having problem to increase the ram but that's the solution Azure is offering but i'm not happy with that.
Whenever applications going down, will clear all the sessions so all the DB connections will be closed and DB will start responding. I know this is not a solution, but we are not able to find a better solutions than this.
I'm wondering how the applications are handling big traffic, i'm little confused. I need some work around to handle this scenarios. Again i'm not looking a solutions straight forward, i'm looking a best practices to handle this situations.
