VMworld 2013: Practical Real World Reporting with vCenter Operations
1 like1,058 views
The document discusses practical real-world reporting using vCenter Operations, focusing on storage analysis and optimization methods. It outlines the importance of utilizing metrics to effectively size storage and report on infrastructure utilization, risk, and waste. The presentation includes strategies for generating custom reports, optimizing performance, and forecasting future growth in IT environments.
VMworld 2013: Practical Real World Reporting with vCenter Operations
- 1. Practical Real World Reporting with vCenter Operations Tom Findling, VMware Monica Sharma, VMware VCM5009 #VCM5009
- 2. 2 Topics: Real World Reporting with vCenter Operations Today & Roadmap Analyze to Size Storage Analyze & Optimize How-to Video 1 2 43
- 3. 3 How to Report Using vCenter Operations Today? • Optimize Infrastructure • Identify & Analyze Resources at Risk • Visualize Top N, Trends & Bottlenecks • Out of box • Views: Capacity & Performance • Export as Report • Custom • Create Custom Heatmaps, • Create, Clone, Edit Dashboards • Import Custom Report Templates Out of box Custom Benefits Description
- 4. 4 Roadmap – 2014: Report Builder and Canned Report Packs • Report across SDDC & Hybrid • Improved Operational Efficiency • Custom Report Builder • Any resource, any metric • 25+ widgets • Drag and drop to create • Out of box Report Packs • Operations Teams • vSphere & Exec Teams • Application or LOB owners Benefit Description
- 5. 5 Topics: Real World Reporting with vCenter Operations Today & Roadmap Analyze to Size Storage Analyze & Optimize Resources for you to download
- 6. 6 How many of you have been asked to size a storage array over the past 2 years? Audience Poll Question
- 7. 7 Most Common Challenges in Sizing Storage Oversized Array Way more powerful than what is really needed Early provision for future Add additional buffers for performance Undersized Array Performance problems from day zero Lack of visibility into future workload demand Unbalanced Array Capacity Waste of money Powerful backend that can’t be utilized Add more disks despite enough capacity? 1 2 3
- 8. 8 What Data Do We Need to Size Storage? Clusters Storage Array Frontend ports – e.g 2 per Cluster Storage Array Backend Storage Frontend: • Total IOPs & Throughput demand for each each cluster? • What is the bottleneck of the frontend ports? Storage Backend: • Total IOPs & Throughput demand for current environment? • IOPs & Throughput my future array needs to support ?
- 9. 9 Methodology to Correctly Size Storage Get the Right Metrics Size Backend Size Frontend Analyze & Optimize Peak IOPs 1 2 30 Download after this session: • Step by Step Video for above
- 10. 10 Get the Right Metrics Metrics to Interpret IOPs (Peak, Average), Read IOPs(Peak), Write IOPs (Peak) Disk Read/Write rates(Peak) Total Disk Usage rates(Peak, Average) Time Range At least 30 days
- 11. 11 Analyze IOPs to Size Backend Peak IOPs: 73,944 IOPs 3 key questions and what to measure/analyze: What is the total IOPs across my environment? What are the Read and Write IOPs? What is the Read and Write throughput? disk|Commands per second disk|Reads per second disk|Writes per second disk|Read rate(Kbps) disk|Write rate(Kbps)
- 12. 12 Analyze IOPs to Size Frontend 3 key questions and what to measure/analyze: Analyze IOPs per cluster(s) Metrics: disk|Commands per second What is the total IOPs across my cluster? What are the Read and Write IOPs? What is the Read and Write throughput? Cluster Time period Average IOPS Peak IOPS Peak Read IOPs Peak Write IOPs Prod 1 30 days 499 2053 1908 389 Prod 2 30 days 1474 10,572 9876 878 SQL Cluster 30 Days 2818 13755 870 1288 DMZ Cluster 30 Days 76 990 78 882
- 13. 13 Compare Avg & Peak IOPs Avg & Peak close to each other Use Peak for sizing Avg & Peak far apart One Time Peak Ignore Cyclical Peak Environment wide activity – AV, Backup Spread Peak over time 1 or more VMs Isolate VMs causing Peaks Analyze & Optimize Peaks 1 2 3 Optimize Peak
- 14. 14 Demo – Analyze & Optimize Peaks
- 15. 15 Demo – Analyze & Optimize Peaks
- 16. 16 Gain True ROI from Optimizing Peak IOPs Before IOPs analysis, peak IOPs = 73,944 Peak found to be an Anti-Virus scan. IT Admin spread Anti virus scan across VMs to reduce peak IOPs Peak IOPs reduced to = 49,303 Savings = 73,944 - 49,343 = 24,601(33%) ~= $300,000 in Enterprise Storage
- 17. 17 Become an Expert Yourself: How-to-video on VMware Blogs More will be coming - Keep reading vC Ops Tech Tips! http://tinyurl.com/mb s78o4
- 18. 18 Topics: Real World Reporting with vCenter Operations Today & Roadmap Analyze to Size Storage Analyze & Optimize How-to Videos
- 19. 19 In the last 6 months, how many of you had to report on Infrastructure Utilization & Unused Capacity? Audience Poll Question
- 20. 20 Why is it Important to Report on Utilization, Risk & Waste ? Am I meeting my SLAs? How well am I utilizing my infrastructure? What does growth look like & do I have enough? • Reporting on Demand, Utilization & Risk • Capacity bottlenecks now or in the future? • Capacity available, used, remaining? Monitor Demand, Utilization, Risk Reduce Operational Risk to Ensure Performance by 1 • Reporting on Waste to Optimize • How to identify unused reclaimable resources? • How to optimize under utilized resources safely? Increase Utilization & Realize Savings by 2 • Reporting on growth & forecast • What has been VM growth trend? • What has been my burn rate? Do I have enough? Forecast for future growth by 3
- 21. 21 Get the Right Metrics 16 GB- Total Allocated Capacity 2GB -What VM did not get (Contention) 8GB - What the VM got(Usage) SQL VM 10GB- What the VM wants(Demand) • Demand is What the VM wants: Physical resources an object might consume w/o constraints • Demand = Usage (what VM gets) + Contention (What VM does not get) • Allocation – Amount of a resource that the user configures • Use Demand for capacity & performance if Demand > Entitlement • May have performance issues • May be undersized (‘Stressed’) • Use Demand vs Consumed for Memory Buffer The most a VM can get (Entitlement)
- 22. 22 Right Metrics for Right Reporting – Where Do I Find Them?0 Available in Product Documentation for Custom Reporting
- 23. 23 Report on Demand, Utilization & Risk1 VM Growth Infra Burn Rate Capacity Risk
- 24. 24 What Has Been My VM Growth Trend? vC Ops vSphere UI Planning Vm Capacity View vC Ops Custom UI->VM Count & Trend
- 25. 25 What Has Been My Infrastructure Utilization Trend? Visuals: Breakdown by cluster to view Actual Demand by Clusters Metrics: Use Usable Capacity vs Total Capacity for Planning decisions (includes Buffers)
- 26. 26 How Well Is My Infrastructure Utilized Today? Cluster Usage vs Total Which Clusters are undersized? Which Clusters are Under-utilized?
- 27. 27 Report Capacity Risk Based on Your ‘Knobs’ Flag risk @ 60% Over-commit CPU 2:1 Don’t over-commit Memory Size for peak usage Higher buffers Enable Alerts Business period preference set Desired Consolidation Ratio Flag risk @ 85% Over-commit CPU 4:1 Over-commit Memory 20% Acceptable stress =20% Lower buffers Disable Alerts No Business period preferences No upper limit on Consolidation Ratio Production Policy Test-Dev Policy
- 28. 28 Which Clusters Are at Capacity Risk & Why? Which clusters are at Capacity Risk? Compare Allocation to Actual Demand Why? - Out of Capacity? - Will run out soon? - Under-Sized?
- 29. 29 Which Datastores Are at Capacity Risk & Why? Datastores at capacity risk –color coded Which VMs Causing most waste?
- 30. 30 Report for Line of Business VMs Resources used by all my VMs VM growth Trend Under-sized VMs (Stressed) VMs at Capacity risk in near future
- 31. 31 Report on Top N VMs with Capacity Risk VMs out of Capacity? Undersized VMs? VMs out of Guest FS? VMs running out of capacity soon?
- 32. 32 Report to VM Owner – Right-sizing (Over-sized) VMs Top N Over-sized VMs Top N by Memory Usage Top N by CPU Usage Trend Memory Usage Trend CPU Usage
- 33. 33 Conclusion & Q/A vCenter Operations Enables You to Improve Operations Visibility & Maximize your IT Investment Q / A
- 34. 34 Other VMware Activities Related to This Session HOL: HOL-SDC-1301 Applied Cloud Operations Group Discussions: VCM1002-GD, VCM1004-GD Cloud Operations with Hicham Mourad or Sam McBride
- 35. THANK YOU
- 37. Practical Real World Reporting with vCenter Operations Tom Findling, VMware Monica Sharma, VMware VCM5009 #VCM5009