Traditionally, identifying and remedying performance problems resulting from application deployments has been a slow, reactive process. Tools exist which report on application changes and problems after they occur, but how do you prevent your next performance issues before they even begins?
Join Correlsense and dbMaestro for an online seminar outlining the crucial strategies for continuous performance testing and monitoring. We will discuss:
-Limitations of traditional strategies for application deployments
-Best practices for eliminating the risks of application changes
-Solutions for proactive application performance monitoring and database change management
Preventing the Next Deployment Issue with Continuous Performance Testing and Monitoring
1. Preventing the Next Deployment Issue with
Continuous Performance Testing and Monitoring
December 13, 2012
Tom Batchelor, Correlsense
Uri Margalit, dbMaestro
2. Agenda
1. History of APM
2. Steps to Proactive Monitoring
3. The Need for Joint APM/DCM
4. Summary/Q&A
6. In the beginning…
• End users would call IT/Help Desk
• IT would try to simulate problem
• No insight into end user experience
7. Evolution to on-demand traces
• Turn on in response to a problem
• Doesn’t help us discover or predict problems
• Still reliant on end users calling in
8. Evolution to 24x7 monitoring
• Capture end user data
• See all tiers
• Measure SLAs 24x7 for all users
9. But today’s solutions are still reactive!
• How do you prevent performance issues before
they exist?
Lack end-to-end performance management view
“CPU and “PING works, “SAN has low “Plenty of Storage &
Memory are OK” Temp is OK” utilization” No hardware failures”
Application Database X
Server Server Fabric
Storage Target
Application Server SAN
Performance SW Mgmt. SW Mgmt. SW
Array Mgmt. SW
SRM Tools
“Plenty of
capacity”
11. Proactive management- know before
the users
Topology map isolates the
infrastructure components
involved
SLA Analysis points Ticket is escalated to right application team
team drills down
to the bottleneck Identifies faulty
area. method
App/ops team provides
workaround/fix.
shows the locations Change impact analysis
that are affected and proves issue has been
user in apps resolved
- detects
degradation in SLA Ticket closed
- Opens ticket
16. Track all requests through all hops
• Track all user requests through all components (not
just Java and .NET)
• Track a single end user across entire stack
• Apache, OC4J, and Database
Applet
Apache
OC4J
Forms Runtime
Specific SQLs
18. Proactive monitoring benefits
• Increase operational efficiency
– Avoid application `brown-outs` and slowdowns
– Reduce staff and time to resolution
• Save on hidden costs
– Fix problems before they become really expensive
– Solve problems in testing before going live
20. Poll
• Which of these do you consider as the biggest
benefit from Database Change Management?
– Team Collaboration
– Change Policy Enforcement
– Development Process Management
– Merge & Deploy Automation
– Preventing Next Performance Issue
21. What is DCM?
• DCM – Database Change Management
• Part of ALM (Application Lifecycle Management)
solutions
• Foundation of Agile & CI
SCM DCM Business Conflict
Automated Continuous
Native Database Req. Code
Deployment Integration
Code Code Integration Resolver
22. The database tier
• Database is major part of the application
– Schema Structure
– PL/SQL Code
– Lookup Content
• Database is a
central resource
• Business data
must be saved
23. Need for DCM
• Lack of order in database development
• No visibility
• “Out of Process” changes
• Not having automated tasks
• Problems in version releases
• Responding slowly to changes in requirements
24. Benefits of DCM – development
• Database changes repository
• Following SCM methods (Check-Out/Check-In)
• All changes are documented
• Control Who can do What, Where, When & Why
25. Benefits of DCM - deployment
• Integrated deployment engine
• Business level audit
• Roles & responsibilities enforcement
28. Summary
• Today’s solutions are still reactive
• Proactive monitoring strategies
mitigate the risk of change
• dbMaestro and Correlsense offer
a joint solution for preventing the
next performance issue
behavior learning and predictive analysis doesn’t work for dynamic environments. Products can’t “learn” enough in order to understand the patterns, as things change all the time. Gartner say so as well. our way of doing that is sensing that something is wrong by being able to see %SLA start to decrease and immediately point the problem even if only a very small fraction of transactions are slow. Quote from FIBI: “We use SharePath to sense that something is about to go wrong and be proactive about it”