The Color of Money - Priortization of User Tasks
Download as PPT, PDF0 likes496 views
The document outlines an agile technique for prioritizing performance analysis tasks among various technical roles, emphasizing task importance and frequency to improve efficiency. It discusses how performance analysts manage their tasks and the allocation of resources to address performance issues. A simulated budget exercise illustrates the prioritization of tasks based on customer needs and importance.
The Color of Money - Priortization of User Tasks
- 1. drive UPA The Color of Money - An Agile Technique for Prioritization of Requirements Len Conte Feb 25, 2004
- 2. Big Picture – Target User • Consumer Performance Engineer, DBA, IT Manager • Performance Manager Manager Technical Infrastructure, Technical Services Manager Performance Analyst Technical Analyst, System Administrator, System Architect, Level 2-3 Support • Capacity Planner Capacity Planner, Capacity Analyst, System Analyst, Level 1 Support 2 February 5, 2013
- 3. Big Picture 50,000 foot view of the user’s work • Define all the work involved with the user’s job • Categorize all the tasks • Rank all the tasks › Frequency - H High, M Medium , L Low › Importance - A High, B Medium, L Low • Scope the system 3 February 5, 2013
- 4. “The Big Picture” Performance Analysts Analysis Manage the Performance Performance Problem Analyze On Demand Continuous Baselining Knowledgebase Performance Issues Health Check Health Check Tool Administration Process Management Capacity Planning & Operation Problem Process EPA Process On Demand Continuous Tool Eval, and Use Tool Mangement Management Management Capacity Planning Capacity Planning 4 February 5, 2013
- 5. Where they spend their time Big Picture - Performance Analyst Tool Administration & Operation 1.0 Tool Eval, and Use Tool Management 1.2 1.1 Administrate DB Talk to BMC support Collect performance Write scripts to space for historical Populate analyze Install software to solve product data ensure reliability database problems 6 1 H A L A 10 L A 14 L B 18 2 H A 1 M A Run collectors in Learn how to use Get data that did not development and QA Apply patches Troubleshoot install transfer performance products environments 3 H A L B H A H A H A 7 11 15 2 Load test.. validate Maintenance of data consistency in Find out why collect Evaluate performance Find out why data presentation layer test environments on agent did not run products didn't collect demand 4 M A H A L A H A 3 H A 8 12 16 Place sheets side by 1 2 3 4 5 Validate Production Automate daily data Set-up silent install data collection Hardware table Frequency of Task Importance of Task collection and for 3rd party worked and data is maintenance H High A High processing distribution valid a 9 M Medium B Medium 5 H A L A 13 M B L C 5 February 5, 2013 17 L C Low Low
- 6. Tasks analysis meeting What our experts did: › Add any missing “Analyst” activities Several key areas missing: Benchmark / Loadtesting › Prioritize the tasks from their perspectives Frequency & Importance (Red/Yellow/Green) › Decide what tasks BMC should focus on We gave them $1,000 in fake $100 bills to spend (Top 10)!! 6 February 5, 2013
- 7. Where they spent the money Task $$ Frequency Importance Gather other (non-BMC) metrics 700 Identify performance problems / exceptions 700 Correlate compare computer metrics with other data sources &500 correlation (SAS) Workload characterization (global library) 500 Analyze performance problems 400 Proactively search for known performance issues 400 Create summarized Performance view / End to end 400 Observe and assist in benchmarking exercises 400 Feed load test information into Predict 400 Not Assigned Not Assigned Majority of customers (4 out of 7) 63% of4400 $$ Identify component of systems that are bottlenecks 300 Compare actual usage with business measure usage/metrics 300 Identify the category of performance problem 200 Provide ad-hoc reports as required 200 Disseminate performance data / graphs 200 Not Assigned 80/20 cut line 80% 5600 Solve performance problems & make recommendations 100 Report on performance problems & status 100 Analyze Load test (Stress test) performance analysis 100 Assets and configuration verification 100 Identify trends & typical system behavior 100 Note: Blue text denotes customer Identify SLA deviations 100 added task Identify data profiles trends & typical system behavior 100 Compare actual usage with forecast 100 Establish performance (SLA) baselines 100 Relate systems & applications metrics to business 100 Determine threshold settings and correlate events for alerting 100 Make threshold setting recommendations 100 Define application process flow 100 Predict the performance of an un-tested configuration 100 20% 1400 7 February 5, 2013
Editor's Notes
- #8: Red red