Capacity usage trends are calculated based on historical data. By default, the longest time period taken into account for calculating the capacity forecast is 180 days. The more historical data up to 180 days are available, the more precise is the calculated forecast. Forecast calculation methods include peak calculation and average calculation, which can be set globally or customized for individual objects. Capacity forecasting is available for nodes, interfaces, and volumes that meet certain requirements, such as being managed in NPM and having enough historical data in the database. References: Monitor capacity usage trends on the network and forecast capacity issues in NPM, Forecast capacity for nodes, interfaces, or volumes in NPM, Real-Time Network Monitoring Tool

Response time is the amount of time it takes for a service request to return a response. It is a critical SLA performance metric for a hosted data center service, as it directly affects the user experience and satisfaction.
High response time can indicate network congestion, server overload, or other performance issues that can degrade the quality of the service. By tracking response time, you can monitor the availability and reliability of the service, identify and troubleshoot potential problems, and ensure that the service meets the SLA expectations and standards. References: How to measure your SLA: 5 Metrics you should be Monitoring and Reporting, Data Center SLA Frequently Offer a 100% Uptime Guarantee, How To Measure SLA Performance in 4 Steps (With Tips).

High Traffic Utilization alerts are triggered when the interface's received or transmitted percent utilization exceeds a specified threshold. However, this may not reflect the actual network congestion, as traffic spikes can occur momentarily and then subside. To avoid false alerts, you can configure the alert to trigger only when the traffic utilization remains high for a sustained period, such as 5 minutes or more. This way, you can filter out the transient traffic peaks and focus on the persistent issues that affect the network performance.
References:
How do I create a high transmit percent utilization alert to monitor interface bandwidth Trouble with High receive percent utilization alert How alerts work