With the release of the Autonomous Health Framework (AHF) 22.3 version, Oracle introduced the AHF Insights report, which added the capability to get a bird’s eye view of the system and use the data that ahf inherently collects for “deeper diagnostic insights”. This report can not only perform a health check of your system but also help with triage and root cause analysis. Presently this is only available on Exadata, and there are plans to release this for non-Exadata as well.
Generating the Insights Report
Once you have installed or upgraded AHF to the 22.3 version, you can use the “ahf analysis create” command to create the AHF Insights report. Although generating the report is straightforward, you must be mindful of a few things. You can choose a particular time window or create the report for the last x hours. It will be generated for the previous 2 hours if you do not specify the time window.
- Specifying the window: You can choose a 12-hour window using the –from and –to options.
$ ahf analysis create --type insights --from "2022-01-05 10:00:00" --to "2022-01-06 10:00:00"
Starting analysis and collecting data for insights
The time difference should not be more than 12 hour(s)
or
$ ahf analysis create --type insights --from "2022-01-05 10:00:00" --to "2022-01-05 20:00:00"
Starting analysis and collecting data for insights
Collecting data for AHF Insights (This may take a few minutes per node)
AHF Insights report is being generated from 2022-01-05T10:00:00 to 2022-01-05T20:00:00
.....
- Although the documentation states that you can use the –last option to generate the AHF Insights report for the past 12 hours, it doesn’t appear to be the case. 1 to 4 hours is the range for the last command.
$ ahf analysis create --type insights --last 5h
Argument value error for --last: value must be within 1-4 range
or
$ ahf analysis create --type insights --last 4h
Starting analysis and collecting data for insights
Collecting data for AHF Insights (This may take a few minutes per node)
AHF Insights report is being generated for the last 4hReport
The AHF Insights report has two major sections – System Topology & Insights.
System Topology
Under the System Topology section, you get a bird’s eye view of your entire system. This section includes your cluster information at a glance, including the number of Nodes, the version information, the databases running from each home, cluster resources, asm diskgroups and much more. It also has a neat option to copy this information as text.
One of the benefits of the AHF Insights report is that it makes the configuration information extremely accessible. Using a single report, you can go from looking at the system topology to finding the configuration of the Storage Cell Servers, as shown below.
Insights
Timeline is the section where the AHF Insights report goes up a gear and uses Intelligence to detect patterns from the events seen in your logs, e.g. alert logs, and prepare a timeline. Unfortunately, my report didn’t have any information errors during the timeframe.
Yet another instrumental component of the AHF Insights report is the comparison reports for kernel parameters and rpms across the database servers. As you can note in the figure below, there is a switch to only list the properties that differ, or you can choose to filter the parameter you want to review
This can be handy in reviewing things like kernel limits or huge pages across nodes.
There is a separate section to display information about database initialization parameters. There is a comparison toggle available here as well; however, it would have been great if we could compare the parameter values across the instances of the same database if RAC
System Change
“System Change” is again a highly convenient feature of the AHF Insights report, where you see your system change history, displaying the original value/state and change made. The changes are categorized & tagged, making it easier to filter.
Those familiar with orachk/exachk reports will find the familiar Best Practices section displaying the critical or best practice findings, which can be filtered based on multiple categories.
In the image below, from under the “Operating System Issues” section, one can navigate to the problem, which in this case was high Disk Average wait time, and then further drill down to get the exact metric with the timestamp and device name for the server.
So as you see, this compelling report can help improve your efficiency 10 folds as a database administrator and help you in numerous scenarios. I can’t wait for Ahf Insights to be released for non-engineered platforms!