The Analysis Tab allows users to graph certain metrics over time, view historical values and correlate the data with issues raised by Indeni. Indeni collects data from network elements and it stores the data in a time series data store. A Time Series is a series of numeric data points of some particular metric over time. A Metric is any particular piece of data that you wish to track over time.
For the purpose of graphing, it would only make sense to graph data that is numerical in value and the value changes over time. For example, CPU utilization, number of concurrent sessions, etc. This is the reason why only certain metrics are graphable. Indeni keeps up to a year’s worth of historical data, providing you with means to efficiently query your data.
As some metrics are polled more frequently than others, the Indeni system provides different polling time windows depending on the metric.
Metrics polled at high frequencies generate large datasets. To allow you to make sense of the information, the maximum number of values displayed for them in a single query needs to be limited.
- Query timeframes of up 24 hours will yield datasets showing a maximum of 1440 elements (1 data point per minute)
- Query timeframes of up 7 days will yield datasets showing a maximum of 2016 elements (5 data points per hour)
- Query timeframes of up 30 days will yield datasets showing a maximum of 1440 elements (2 data points per hour)
- Query timeframes of up 365 days will yield datasets showing a maximum of 1440 elements (4 data points per day)
An example of why you would use the analysis tab. If you have a pair of devices that form a cluster for high availability or load balancing, it often helps to visualize and compare their resources such as CPU usage for a particular time period. If the configuration is active/standby, you would expect the standby device to have low CPU utilization
CP-2 is the standby gateway and CP-1 the active gateway. In this case, CP-2 is showing high CPU spikes that do not appear to be normal for a standby device that is idle for the most part. Next, you may want to add more metrics for correlation and further analysis. You can add load average, concurrent connections, memory usage, just to name a few. The analysis tab allows you to correlate multiple metrics across multiple devices.