[prchaudhari_007]

We are using a similar strategy. We have 4 different node groups 2 with on demand instance types 2 with spot instance types It has been working really well for us. We are doing some further optimisations to reduce the costs and improve the performance of our stack. Also, we have jenkins scheduled jobs which kills the staging infra every evening and starts it up in morning. We also keep the staging completely down over the weekends.