Efficient Support System
Beth McLamb
Anurag Saini
This session will provide guidelines to
increase the
efficiency
of an OTM support
Having the ability to meet, or exceed, agreed upon Service
Level Agreements between system administrators and customers, while providing
ongoing value to the organization.
What is an
efficient
support
system?
Where do we start?
• Define which components of OTM are supported and release levels (OTM, GTM, FTI)
• Identify external systems implemented (pcmiler, milemaker, rateware, etc.)
• Document integration points and responsible groups for each
• Gather environment documentation – server details, operating system information, patches
Define Components Required
Infrastructure Support OTM Technical Support OTM Application Support OTM Application EnhancementsInfrastructure Support
Infrastructure Support OTM Technical Support OTM Application Support OTM Application Enhancements • Hardware • Network • Storage • Operating SystemOTM Technical Support
Infrastructure Support OTM Technical Support OTM Application Support OTM Application Enhancements • OTM Patches/Upgrades • Property settings • DatabaseOTM Application Support
Infrastructure Support OTM Technical Support OTM Application Support OTM Application Enhancements • Software issues • Software performanceOTM Application Enhancements
Infrastructure Support OTM Technical Support OTM Application Support OTM Application Enhancements • Software improvementsThere must be harmony
Customer Expectations Operating Level Agreement (OLA) Service Level Agreement (SLA) • Internal agreement between OTM Support Groups • Agreement between OTM Support Groups and OTM End User / CustomerExample – environment refresh
Lead time for request Time required to execute request Commitment to fulfill request
Components of an SLA/OLA
• Quantitative performance standards
• Agreed upon severity levels and
turnaround
Establish responsibilities
• Create a RACI Matrix to identify roles and assignment of cross-functional responsibilities
– RACI represents:
• R – Responsible – who is assigned to do the work • A – Accountable – who owns the success of the work • C – Consulted – who can give valuable input
Support Structure
• Each organization is different and responsibilities may be assigned differently between groups.
• Some tasks may be managed by external groups or everything may be internal.
• End users should have a single point of contact and the assignment of tasks should be decided by the support organization.
Engaging Support – single point of contact
Issue or Work Request
Screen Enhancement Missing Transaction Environment Refresh
Defining issues
• Implement a solid issue reporting/tracking tool
• Educate those entering the issue to supply adequate detail
• Identify request type – incident, work, enhancement • Determine severity and response time
Sample incident response matrix
Priority Business Impact Description Response Time Resolution Time Severity 1 System DownSystem is inaccessible. 30 minutes 4 hours
Severity 2 High Impact A condition or issue that has an adverse impact on operations until unresolved.
1 Hour 8 hours
Severity 3 Degraded Operation
Question concerning product
performance, or an intermittent low impact issue.
2 Hours 24 hours
Severity 4 Minimal Impact
A question concerning general utilization/implementation
24 Hours As agreed upon
People – your most valuable resource!
What makes up a
Support
Identify the skillset
• Based on your support definition – gather the right people
• Bring support team in early during your implementation to allow for overlap and knowledge transfer
• Invest in high quality training
• Don’t allow application support and enhancement teams to work in silos
Crossover between support and
enhancement groups
OTM Support
OTM
Build your toolset
• Create meaningful documentation with detailed issue resolution
• Follow a well documented migration strategy to keep environments in sync
• Automate mundane tasks to keep the team challenged • Reward your team for innovative ideas and adding value
OTM is
powerful –
avoid common
mistakes
Whether a new implementation or one that has been live for years, don’t allow these to rob precious time from your support
Agents – power with caution
• Recommendation: One Event One Agent. Many agents can be triggered by a single lifetime event. These agents work in parallel.
• Badly configured agents can be a performance hit in the long run. Review frequently as volumes environment
changes.
• Agents run as serial workflow on the application server. One action does not start executing until its predecessor has completed If an error occurs within an agent,
Recurrent deviations
• Create a process to handle “recurrent deviations”. Don’t continue to fix the problem. Find a proper way of
reporting “recurrent issues” and fix the root cause
• Implement a short term solution to keep the business running, but a also have a defined RCA process and plan to resolve the issue permanently
Modifying public data
• Do not modify public / default data or objects as they could get overwritten with a patch or upgrade
• Place objects in a custom folder to remove the dependency on upgrades
Purging
• Insufficient (or non-existent) purging of transmissions • Insufficient (or non-existent) shipments/orders/invoices
OTM Logging
• Unnecessary logging impacts performance perspective and may hinder analysis of the logs when issues do
arise
• Adhoc logs enabled could cause perf issues as all of the logging enabled is logging for the entire domain and not just that user. Change these to user logs.
Inactivate Rates, users
• Inactivating expired rate records
Modifications to DB Schema
• The OTM DB Schema is complicated and consists of over 1600 tables. This requires an experienced skillset to manage.
• Many OTM customers end up adding or deleting indexes to optimize performance.
Remember to document them, as they may get re-created when OTM patches are applied. ( Refer to Doc ID 762939.1)
– "It should be noted that these indexes have to be maintained going forward or a request can be submitted to have the indexes added in the OTM Core product in a future release. Indexes will be added only if it is beneficial for the application
performance as a whole and not for a specific implementation“
• Never allow modifications of packages or custom packages to go into GLOGOWNER, REPORTOWNER, or any other OTM built-in schemas, create a custom schema for this
Working with Oracle Support
• Gather all details for the issue in simple language • Gather steps to recreate the issue and what was
happening when it occurred
• Run qdlogs script and attach output to the ticket
• Run the OTM analyzer script – My Oracle Support note 1579683.1