Sunday, September 30, 2012

EM 12c

Kellyn Pot'vin presenting on EM12c - full house!

Why monitor? To  be proactive. Keep the users from becoming annoying, and keep management in the procurement cycle.

Challenges: White noise, wrong noise, late noise.

Manage incident manager.  Create incident from alerts using rules.  New visuals, new shift.  This is a shock and a learning curve.

To use existing rule sets - don't mess with the existing ones.  Instead disable and create-like.  Several reasons:  they work; they might get overwritten with patches.

Highlight: - the metric thresholds are only set for Warning, not Critical.  If you have incident notification on "Critical only", default results in no messages.

"Metric Extensions" replaces "User Defined Metrics".   Much more flexible, wider scope, build your own library, keep your own stand-alone scripts and fold them in.  Has a 'development lifecycle', so you can test *before* production.

Performance pages have not changed much.  "Pink (and Orange) is not a girl's best friend".   SQL Monitor is very useful - learn it if you don't know it yet. 

ASH Analytics!  Real Time ASH.  Drill into events, SQL, SQL details, over a variety of time frames. Load Map is a new way of showing the load.  Very useful, especially when trying to convince people to address an issue.  Replaces "Top Activity", which will likely go away.

No comments: