|
|
|
The data which is used to generate graphs in trend reports is stored as archives in a round robin database. These archives are of fixed length; when they fill up, incoming data overwrites the oldest entries.
Archives are written to files which are stored on the system’s hard disk. See Backing Up Reports, Trend Data and Saved Configs in Chapter 2: Configuration Fundamentals of the CNA Fundamentals Guide for information about maintaining a backup of the trending database.
The database files store both raw data and calculated averages. (This distinction involves a technicality, since even the raw data represents an average of actual raw data obtained during the hour between snapshots. The distinction is between data as it is recorded in the snapshot--which is “raw” as far as the trending function is concerned--and averages of that stored data as calculated by the trending process.)
Whether graphlines are plotted from raw data or average data is determined in part by how you have configured the trending database using the trending size (deprecated) command, and in part by the time period you specify for the graph.
This relationship among database size, time line and graph should be well understood by anyone who creates trend report definitions.
Note that a very heavily loaded CNA system may take some time to complete the trending functions. If this time is longer than the snapshot frequency you have scheduled, then the snapshot will not complete as often as expected. For example, if you have scheduled snapshots for every hour, but trending takes one hour and 10 minutes, then the CNA system will only be able to do a complete snapshot every other hour.
|
|
|