For example, metrics that measure disk activity have a disk-device dimension. Aggregating on the disk-device dimension shows the average disk activity for all disks, where pivoting (splitting) by the disk-device dimension shows the activity for specific disks.

Every metric has a logical instance dimension, which corresponds to the server on which the sample was created. Each metric may have up to three dimensions, which are defined in the metric definition.

For example, the sync-pipe-completed-ops metric has two dimensions, the pipe-name and pipe-result. The pipe-name is the name of the sync pipe as configured for the PingDataSync Server. The pipe-result is one of the following values:

  • exception
  • failed
  • failed-at-resource
  • failed-during-mapping
  • match-multiple-at-dest
  • no-match-at-dest
  • already-exists-at-dest
  • no-change-needed
  • out-of-scope
  • success
  • aborted-by-plugin
  • failed-in-plugin

At each measurement interval for each sync pipe on each PingDataSync Server, there will be a value for each of the pipe-result values. So, for a single PingDataSync Server with two Sync Pipes, pipe-one and pipe-two, the samples generated for each sample period look like the following. The timestamp is constrained to time-only for brevity.

08:15:05, sync-pipe-completed-ops, pipe-one, exception, 1
08:15:05, sync-pipe-completed-ops, pipe-one, failed, 7
08:15:05, sync-pipe-completed-ops, pipe-one, failed-at-resource, 1
08:15:05, sync-pipe-completed-ops, pipe-one, failed-during-mapping, 1
08:15:05, sync-pipe-completed-ops, pipe-one, match-multiple-at-dest, 3
08:15:05, sync-pipe-completed-ops, pipe-one, no-match-at-dest, 0
08:15:05, sync-pipe-completed-ops, pipe-one, already-exists-at-dest, 0
08:15:05, sync-pipe-completed-ops, pipe-one, no-change-needed, 1
08:15:05, sync-pipe-completed-ops, pipe-one, out-of-scope, 1
08:15:05, sync-pipe-completed-ops, pipe-one, success, 125
08:15:05, sync-pipe-completed-ops, pipe-one, aborted-by-plugin, 1
08:15:05, sync-pipe-completed-ops, pipe-one, failed-in-plugin, 0
08:15:05, sync-pipe-completed-ops, pipe-two, exception, 3
08:15:05, sync-pipe-completed-ops, pipe-two, failed, 9
08:15:05, sync-pipe-completed-ops, pipe-two, failed-at-resource, 2
08:15:05, sync-pipe-completed-ops, pipe-two, failed-during-mapping, 1
08:15:05, sync-pipe-completed-ops, pipe-two, match-multiple-at-dest, 4
08:15:05, sync-pipe-completed-ops, pipe-two, no-match-at-dest, 0
08:15:05, sync-pipe-completed-ops, pipe-two, already-exists-at-dest, 0
08:15:05, sync-pipe-completed-ops, pipe-two, no-change-needed, 1
08:15:05, sync-pipe-completed-ops, pipe-two, out-of-scope, 1
08:15:05, sync-pipe-completed-ops, pipe-two, success, 217
08:15:05, sync-pipe-completed-ops, pipe-two, aborted-by-plugin, 1
08:15:05, sync-pipe-completed-ops, pipe-two, failed-in-plugin, 0

Compare how busy pipe-one is to pipe-two by pivoting on pipe-name. This results in the following:

pipe-one 141
pipe-two 239

Pivot by pipe-result, to get a set of counts that show the distribution of the counts of the specific error types, as well as the success and failure. This data provides a quick way of assessing the kinds of problems encountered by the Sync Pipes.

Dimensions provide a way to pivot or aggregate along a metric-specific axis. All metrics have the instance pivot and the time pivot. Metrics that support the histogram statistic can also have a histogram pivot.