AHF Insights

Oracle recently released a new version of its Autonomous Health Framework (AHF). At the time of writing the version is 22.3.1. 
As the development is quite active there, one of the new features is AHF Insights with a nice description: 

AHF Insights provides deeper diagnostic insights into Oracle diagnostic collections collected by AHF diagnostic utilities, Oracle EXAchk, Oracle Trace File Analyzer, Exawatcher, and Cluster Health Monitor.

This sounds promising, so I was tempted to give it a try! 
My first attempt failed: I gave it a chance at my sandbox in my laptop, but right now AHF Insights only work on Engineered Systems. I didn't found this documented somewhere - but ok. The good news: I was told AHF Insights will be available in future versions also for non-Engineered environments. As this is the first release, I can follow the argument here. 

So I had to grab an Engineered System machine, install AHF, let it run for some time and then just run 

ahf analysis create --type insights

I also could be executed for a specific time in the past (with --last n{m|h}) or some range (--from DATETIME --to DATETIME) - But as I was not searching for a specific issue, I took the default 2 Hours. 

The execution took some minutes, but it was faster than an EXAchk. 

The zipfile created has a fileformat like <hostname>_insights_YYYY_MM_DD_HH24_MI_SS - it contains one directory web containing an index.html and a lot of subfolders:

.
└── web
    ├── css
    │   ├── alta
    │   │   └── 11.1.1
    │   │       └── web
    │   │           ├── fonts
    │   │           └── images
    │   │               └── sprites
    │   ├── fonts
    │   ├── images
    │   └── redwood
    │       └── 11.1.1
    │           ├── common
    │           └── web
    │               ├── fonts
    │               └── images
    ├── dynamicHtml
    ├── icons
    ├── js
    │   ├── jsons
    │   │   ├── insights
    │   │   ├── meta
    │   │   └── topology
    │   ├── libs
    │   │   ├── oj
    │   │   │   └── v11.1.1
    │   │   │       └── resources
    │   │   │           └── root
    │   │   └── require
    │   ├── viewModels
    │   └── views
    │       ├── compliance
    │       │   └── js
    │       │       └── input
    │       ├── dbParameter
    │       │   └── js
    │       │       └── input
    │       ├── ganttPage
    │       │   └── js
    │       │       └── input
    │       ├── os
    │       │   └── js
    │       │       └── input
    │       └── rpm
    │           └── js
    │               └── input
    └── log

the welcome.html is all you need to open. In my case the first page looks like this: 

It contains some basic information about 

• the AHF version 
• the time range it covers
• System Topology
• Insights it found

The Cluster information contains basic details about the cluster itself, it's resources and the ASM disks. 
The summary information has anCopy as text button which generates a nice output: 

========================================================================================================================
Area                           | Dimension                      | Value
========================================================================================================================
System                         | Type                           | Exadata
System                         | Total Node Count               | 2
Cluster                        | Node Count                     | 1
Cluster                        | GI Version                     | 19.17.0.0.0
Cluster                        | Timezone                       | Europe/
Cluster                        | Cluster Name                   | 
Cluster                        | CRS Home                       | /u01/app/19.0.0.0/grid
Databases                      | Count                          | 1
Databases                      | Database Home                  | /u02/
Database Servers               | Count                          | 2
Database Servers               | Hardware Model                 | Standard PC (Q35 + ICH9, 2009)
Database Servers               | Image Version                  | 21.2.13.0.0.220602
Database Servers               | Operating System               | Linux x86_64
Database Servers               | Operating System Version       | 4.14.35-2047.511.5.5.1.el7uek.x86_64
========================================================================================================================

Unfortunately this feature isn't present on all the pages. 

The Cluster Resources and ASM disks just contain information about all OCR resources and the ASM diskgroups. 

The Databases Topology information provides some basic information about the CDBs and their PDBs, together with some very basic instance information like cpu_count, sga_target, pga_target, pga_limit and processes. 

On the Insights section, those with some findings worth to check are highlighted in red. On my system, there are 4 such sections. I will show some of them as examples: 

The Timeline gives a graph of all relevant events, in my case there are only 2 types of events on one hostname.  Below there are filters which can be used if many events appear on the timeline. Also a list of Timestamp, Type, Event, Hostname and Description of each (filtered) event is shown. 

If you move the mouse to the right top of the chart, a list of icons becomes visible for an interactive interaction with the graph:
This can be used in all charts in the Insights.

In the Operating System Issues Page, a lot of detailed information about CPU, Memory, IO, Network, and Processes is shown, both as Charts in Metrics and Grouped Details in the Reports section.  
A nice detail in Process Aggregation is the aggregation on instance basis - so I can check the aggregated memory, processes, etc. for each instance. 

In the Best Practices a list of all the checks, their findings and recommendations is given. 
I always be careful with "Best Practices" - in this particular case the underscore parameters were given by Oracle support in a specific service request. 
But of course, you can see them as recommendations, either follow them or at least document the specific reason you did NOT follow them for a given purpose. 

In the last section, Recommended Software, you can see the early stage of the software. 
For me, 19.17.0.0.0 is close enough to 19.17.0.0.221018.
Maybe in a future release, Oracle will agree ;) 

At all, I can highly recommend to use AHF Insights whenever you need to investigate an issue which might have a wider scope than one single SQL or instance. Even as a first release, it can provide a lot of useful information in a nice and well structured report. 

TFA reports full dba_feature_usage_statistics to MyOracleSupport

Those who deal with MyOracleSupport (MOS) regularly will have noticed their demand for Trace File Analyzer (TFA) diagnostic collections.
Few will know they are also providing a full content of their DBs DBA_FEATURE_USAGE_STATISTICS view; at least when running the parameters

tfactl diagcollect -database <DB_UNIQUE_NAME>

The report calls tfa_home/resources/sql/db_feature_usage.sql and it's result is stored in the <hostname>.tfa_<timestamp>.zip file.

I did not find any mention about this behavior in the documentation, so it might be worth to post about it.

This is at least true for version 18.3.0.0.0 (latest version available at 2018-09-04)

oratop with TFA 12.2.1.3.0

We are facing an issue where Oracles Trace File Analyzer creates high load on some servers.
During the discussion with Oracle Support I installed the latest version of TFA which is 12.2.1.3.0 right now.
With this version, oratop - which oracle calls

oratop is a utility similar to the unix OS utility top which gives a overview of database performance and can be used in combination with top to get a more complete overview of system performance

did not work as expected but quit with

Oratop successfully tested. 
ohome /appl/oracle/product/rdbms_12102_Oct17BPJa 
ouser oracle 
osid TTT0091 
db_running 1 
running_local 1 
oversion 12.1.0.2.0 

After another SR , it was clear to be Bug 27103547 - ORATOP DOES NOT WORK IN TFA 12.2.1.3.0
The solution is simple:
In $GRID_HOME/tfa/`hostname`/tfa_home/bin/common/tfactlshare.pm
edit somewhere around line 9010

# Farm called 
if ( $tcase ) { 
if ( length $ohome && length $ouser && length $osid && 
$db_running && length $oversion ) { 
print "Oratop successfully tested.\n"; 
print "ohome $ohome\n"; 
print "ouser $ouser\n"; 
print "osid $osid\n"; 
print "db_running $db_running\n"; 
print "running_local $running_local\n"; 
print "oversion $oversion\n"; 
# exit 0; 
} else { 
print "Oratop failed.\n"; 
print "ohome $ohome\n"; 
print "ouser $ouser\n"; 
print "osid $osid\n"; 
print "db_running $db_running\n"; 
print "running_local $running_local\n"; 
print "oversion $oversion\n"; 
exit 1; 
} 
} # end if $tcase 

So the exit 0; is not executed anymore.
Now oratop works as expected - even with TFA 12.2.1.3.0

---

Update: 18-Dec-2017
A new version 12.2.1.3.1 is available - this Bug is fixed there.
thank you Sandesh Rao for informing me.

@martinberx the oratop issue which was in TFA 12.1.2.3.0 has been fixed in 12.1.2.3.1 which was just released , thanks for pointing it out

— Sandesh Rao (@sandeshr) December 17, 2017