How do make sure Clarity is "healthy?"

Document created by Thang on Feb 25, 2015
Version 1Show Document
  • View in full screen mode
As a Clarity Administrator, do you use a checklist or some kinds of home-grown guide to manage the more routine aspects of Clarity? If so, what things do you track to make sure your instance of Clarity is running healthy?
I would like to also ask the add-on questions:
1) Who has instrumented Clarity?
2) What tools do you use to monitor Clarity?
3) Are you monitoring all Clarity components (H/W, O/S/, Database, Websphere/Tomcat/Weblogic, services, etc)?
4) Are you doing any event correlation and automated recovery actions?
5) Are your monitoring tools integrated into the Service/help desk?

I am asking these additional questions as Enterprise Monitoring/Help desk integration was my prior field and I have seldom seen a Clarity implemention setup with performance/capacity/availability monitoring.

I look forward to your responses!
Michael Thibault
If you define available and giving continuously best performance you could look at (maintain bench marks) when your start your day

Login, check time to login page to display, start page to appear
Run a report
That will tell if the servers are up and running, accessible and communicating

Job schedule, are jobs completing on schedule, not processing, not scheduled in the past, not waiting eternally.
Scheduled reports, the schedule should be in the future
If any of these check the timeslice last updated date and time to see if bg is running.

If any problems with the above that will tell if the services are up DB, App, Resports/BO services, bg

Check if beacon service is up if used

Are there errors or warnings in App, BG or System, Logs
If there are repeated errors, trace when, who doing what and was that a problem for the user or to the system
Accessing the logs in NSA/CA SA will tell if nsa service is up
Health Check
System Performance tracked using NSA performance tool
Timesheets for the past time periods
Remaining Invalid transactions

Is there a version specific database tuning and maintenance plan and is it followed

CPU Usage for Performance,
Main Memory usage, establish after boot level
Tune java vm memory
Java Memory actual usage in system.caches, establish after boot level
Database size, table size, free disc space
System performance in NSA
Job schedule for level loading of jobs and minimizing overhead data and user session, job logs etc.
Intuitive naming convention to display important jobs on the first page, paused jobs if any next and temporarily, completed jobs at the end.
Use "Do not display data until I filter" setting for portlets
Use query based portlets instead of object lists for objects with large number of custom fields and instances
Personal or site links for frequently accessed pages with long crumb path
Verify that SQL trace is on only when needed
Define a server reboot policy

Define user closing policy, minimize obsolete user foot print, remove OBS associations, rights, group memberships, collaboration manager positions when removing project manager position, removing from collaboration teams and inactivate
User session timeout setting to drop forgotten sessions
Define investment closing,purging and inactivating policy
Number Notifications, are the notification cumulated or deleted when obsolete
Initiated process instances, is there a purging policy
Process Engines, heart beat
Process runs, are all completed and none hanging
No completed jobs or reports in the scheduled list (that is not a log of runs, but schedule for the future
Define time slices for your real need
Monitor the number of collaboration based instances, discussions, notifications etc
Purge Obsolete documents
Long gone timeperiods

License verifications, does the license count match the purchased licensed

Appropriate custom tests for customizations (supported and unsupported

which are some of the things brought up on LinkedIn Clarity groups

Martti K.
It would be great and good for clarity administrators :grin: if CA can enhance the NSA / Application Admin (In future version) considering the above suggestions.......
(i.e GUI in NSA -->App / DB / Process Performance Monitoring in graphical formats).

Hi Paul.
I see two separate questions - I'd like to separate them. We should also pay attention to "theory" and "practice". I'm going to tell you about the latter from my seat, the Clarity/Business Object Admin/Architect/Engineer/Developer & Technical Support team of one.
As a Clarity Administrator, do you use a checklist or some kinds of home-grown guide to manage the more routine aspects of Clarity?
Yes, we have Admin business processes that are documented although not all in a "checklist" fashion. If the process is tight enough to be a checklist, I automate it (e.g. time period roll & other data management jobs).
If so, what things do you track to make sure your instance of Clarity is running healthy?
I read this as "how do you monitor application health?" - let me know if incorrect. My answer is "As an application administrator - I don't".

It's called Value Engineering. We're a small architecture deployment with a technical team of one (me), therefore I can't spend time working on problems I don't have. I trust my DBA team are doing their job ensuring everything database side is humming, network team is doing their piece, desktop team their piece, etc. Application side, I make sure I've deployed the architecture properly, applied a couple old school J2EE tricks (-Xmx1300m -Xms1300m for 32bit Java on Windows), automating what I can and simply ensuring we stay in "safe harbor" as much as we can. We have a high performing highly available Clarity app. The app is "healthy" until proven otherwise - this where exception based reporting comes in.

Exception Reporting:
- We're on Win32, we have Microsoft System Center Operations Manager (MSCOM) monitor the niku services (app, bgs, becon, reports). I get notified via email if any of the services go non responsive.
- I have my email address in for the Admin account that receives all fail notifications from reports and jobs (remove Admin from "success" notifications where possible). This way, I get an email if a report or job fails.
- We have scheduled reports that run out of Business Objects. In CABI 2.x (XIR2) I haven't found the way to be notified of failures. In CABI 3 (XIR3.1) I have seen it & have my email address in here but we're not yet in production so I can't speak to how well this works.
- emails, tickets and drive-bys from users.

When I get out of bed in the morning, I check email. If no 'fail' emails from the systems, I log in real quick to just see the app load. If it does, it's "healthy" until proven otherwise.

Like I mentioned we're "value engineered" on CA's small architecture. This approach probably wont fly on larger implementations - but I'm assuming they can staff for "monitoring health" also. This simply isn't in our budget. In our region, most implementations are 1500 users or less - we're all real similar in operations and this is how they operate too. I'd be curious to hear from others qualified with their implementation architecture, number of users supported and size of support team. I hope we'd see different stories - but will we?
Any idea why this thread is still marked as an "Announcement", interesting it may be, but an "Announcement" it is not! :bashful:

(I've created a link to this thread in the FAQ thread sticky though)
It is not Anouncement any more and I can't see any Anouncements at all.

Martti K.

another_martink wrote:

It is not Anouncement any more and I can't see any Anouncements at all.
When I look at the "Recent posts" view (which is what I usually use) this thread always appears near the top (currently #6 in my list) and it is marked as an announcement.

See image below - the "Announcement" tooltip gets displayed when I hover over the icon next to the post name.

So I currently see 3 urgent posts (all the same 8o ) from Chris then 2 sticky posts cool) and then 5 announcements.

OK thanks, have not looked at it since then and it is still announcement.

Martti K.

This document was generated from the following discussion: How do make sure Clarity is "healthy?"