I elaborate the use case here.
Predict CPU, Memory & disk space failures during transaction processing.
– A scenario would be clients do processing salary to the customers through the bank app (this is my client) occurs every week and the Volume spikes for the main server component and the Messaging
– ie. client sends payment files for processing for its customers every Tuesday. Large number of transaction volume spike causes backlogs and processing delays across the Channels how to rectify it?

Data: I have data for 1 month with time stamp of every fifteen minutes for cpu and memory utilization, scan pages per second, virtual memory utilization and all.

we need to find anomaly and do prediction on volume of transactions also for future period and our prediction should help for predictive maintenance to reduce the down time.

In case of any additional clarification, pl revert me.

Pl suggest some approach/models, Thanks,
Ravi

I am currently working for client he has been made payment to their employees every week and that time their huge volume of transaction files passing through the bank app but it creates backlog and takes time to get delivered for that i need to predict that. I have data of every 15 mins cpu and memory utlization and scan per pages per second and all. How to do prediction using either poison or timeseries. besides i need to dom predict maintenance that usuall downtime will get reduce? Pl help me how to address it and let me know if u require more details on it?

There is a small typo in section Example 1: Poisson is faster, the lamada should be 65 to get a probability of around 4.1%.