All cloud objects are elastic and ephemeral. It is a real problem to understand, analyze and predict their behavior for Cost optimization and Capacity management. The raw data is collected by observability tools, but it is big and messy. We explain how to turn that mess into information.
The essential requirement to do cloud Cost optimization and Capacity management is the system performance data about object such as clusters, containers, serverless objects, databases, and virtual discs. The presentation is to explain and demonstrate how the data should be cleaned by anomaly and change point detection without generating false negatives like seasonality. How that should be aggregated addressing the issue of jumping workload from one cluster to another due to “rehydration”, releases, and failovers. How to summarize the data to avoid sinking in granularity. How to interpret the data to do cost and capacity usage assessments. Finally, how to use that clean, aggregated, and summarized data for Capacity/Cost Planning by using ML/Predictive analytics.
IMPACT 2023 Proceeding Session Video: