Disk space requirements and monitoring intervals

The metrics backend on the monitored servers is responsible for the temporary storage of metric data and is configured to keep a maximum amount of metric history based on log retention policies, which are configured with the dsconfig tool.

The default retention policies define a cap on disk space usage, which in turn determines the amount of metric history retained. If the PingDataMetrics server is stopped for a period of time, the monitored servers should be configured to retain enough metrics history to prevent gaps in data when the PingDataMetrics server restarts. The amount of disk space required for metrics history might also depend on the monitored server’s Stats Collector Plugin settings. In general, 500MB is enough to retain an eight-hour span of metrics history.

The value of the sample-flush-interval property of the monitored server’s metrics backend determines the maximum delay between when a metric is captured and when it can be picked up by the PingDataMetrics server. The flush interval can be set between 15 and 60 seconds, with longer values resulting in less processing load on the PingDataMetrics server. However, this flush interval increases the latency between when the metric was captured and when it becomes visible in a chart or dashboard. Changing the sample-flush-interval attribute to 60 seconds has the PingDataMetrics server keep 2000 minutes of history.

The number of metrics produced per unit of time varies based on the configuration. No formula can compute exact storage required for each hour of history. However, 60MB per hour is a standard estimate.