FSICOST06: How do you track your workload usage cycles?

Financial services workload usage can be cyclical and can have usage spikes during specific days like month-end or quarter-end, or it can be intra-day during specific hours. AWS provides customers with a number of usage monitoring services that can scale your operations up and down as demand conditions require. Monitor cost at an application-level, and a workload-level on a regular basis, and optimize usage of resources and cost.

FSICOST06-BP01 Monitor your workload usage cycle around times of higher and lower utilization (quarter-end, year-end, weekends, and holidays) to identify ways to reduce your costs

You may have workload usage cycles for week-end or month-end, and quarter-end have more usage of resources. In some cases, there could be higher usage due to events like the start of trading hours, holidays shopping, and so on. Monitoring usage and corresponding events are helpful to optimize cost and architecture. You can choose to shutdown unused instances, for example Amazon EC2 servers for development, or QA on Friday, and bring them back up on Monday.

Scale generative AI inference endpoints and vector search infrastructure dynamically with observed diurnal patterns. Pre-warm minimal capacity only for peak trading or batch-report windows, then decay to zero or low-cost tiers during off-hours. Automate these adjustments via scheduled scaling policies or event-driven Lambda functions to minimize idle inference costs.