I was looking at converting most of my collections to RAW instead of TABLE. Use of RAW appears to make most of my reports and their SQL substantially more accessible, I don't have to do some pretty gnarly multiple joins against MGMT@METRIC_CURRENT tables. But in looking at one of your own plug-ins, I was surprised to see that when you do RAW, you don't have any CONDITIONS specified in the default_collection.
Can metrics collected as "RAW" also contain CONDITIONS? For example, for CpuUtilization here, if SimpleMetrics are collected as RAW, would this threshold still work? Here is the metadata snippet:
<Metric NAME="SimpleMetrics" TYPE="RAW" CONFIG="TRUE">
<Display>
<Label NLSID="xxxxx">Simple Metrics</Label>
</Display>
<TableDescriptor TABLE_NAME="MGMT_SXXXXX0_SIMPLE_METRICS">
<ColumnDescriptor NAME="CpuUtilization" COLUMN_NAME="CpuUtilization" TYPE="NUMBER">
<Display>
<Label NLSID="xxxx_cpu_utilization">CPU Utilization %</Label>
</Display>
</ColumnDescriptor>
And the Default Collection snippet:
<CollectionItem NAME="SimpleMetrics">
<Schedule>
<IntervalSchedule INTERVAL="10" TIME_UNIT="Min"/>
</Schedule>
<Condition COLUMN_NAME="CpuUtilization" CRITICAL="90"
WARNING="80" OPERATOR="GE" OCCURRENCES="3" MESSAGE="The CPU Utilization
Percentage is %value%
and has remained above the warning (%warning_threshold%) or
critical (%critical_threshold%) threshold."
MESSAGE_NLSID="cpu_percentage_used_cond"/>
Next, I noticed that your CMDB information tended to be collected as RAW whereas your detailed statistics were collected in a TABLE. Is there an advantage to collecting static information within a table and highly variable information as RAW or was this simply reporting convenience?
Finally, many of highly variable statistics I want to present as time-span graphs. If I collect a stat is RAW, do the MGMT$METRIC_HOURLY (etc...) tables still get populated?
Thanks for the Help
Paul Monday