Hands-on Performance Workshop - The science of performance
Download as PPTX, PDF0 likes897 views
The document outlines a hands-on performance workshop led by Mike Croft, focusing on the science of performance, tools available for performance analysis, and practical experience in data collection and interpretation. Key topics include environment setup, performance factors, data collection methods like garbage collection and thread dumps, and tools for monitoring and analyzing performance data. The workshop aims to enhance participants' knowledge and application of performance analysis tools.
Hands-on Performance Workshop - The science of performance
- 1. Hands-on Performance Workshop The science of performance Mike Croft
- 2. Who Am I? • C2B2 Expert Support consultant • Ex-IBM • Snowboarder • @croft
- 3. Who are you?
- 4. Why are we here? • Hands-on performance! – Hands on what? • See what tools are available • Get experience of using some tools • Learn how to apply knowledge • Why do you want to be here?
- 5. Agenda • Introduction • Environment setup • Performance Overview • Collecting Data – Presentation – Practical • How do we interpret the data? – Tools • Example app analysis
- 6. Environment Setup Hands-on Performance Workshop
- 7. Environment Setup • Download Instructions and follow the first part https://s3-eu-west-1.amazonaws.com/jaxlondon2014/instructions.docx • Make sure all necessary resources are downloaded • Make sure the app is deployed and JMeter can run load through it
- 8. Performance Overview Hands-on Performance Workshop
- 10. Performance Overview Performance Factors • Raw Algorithmic Performance • Resource Limitations – Not enough cpu, disk, memory • Resource Contention – Locks • IO Latency – Network, Disk
- 11. Performance Overview Latency Factors • Network Distance • Network Reliability • Data Size • Operation Granularity • Resource Contention • JVM GC
- 14. Collecting Data Hands-on Performance Workshop
- 15. Collecting Data • Garbage Collection • Verbose GC • Heap size – New size and old size – Before collection – After collection • Pause time
- 16. Collecting Data • Thread dumps – Kill -3 – JStack • Thread state – Wait – Sleep – Blocked – Running • Full stack trace
- 17. Collecting Data • Heap dumps • Entire contents of the heap – Very Large! – Can take time to collect on large heaps • Can auto-dump on OOME
- 18. Interpreting Data Hands-on Performance Workshop
- 19. Interpreting Data • Tools – Live monitoring • VisualVM • MissionControl • JMX – Historical analysis • Memory Analyzer Tool • Flight Recorder (commercial) • Threadlogic • GCViewer
Editor's Notes
- #10: Performance in middleware comes down to: How fast does a single transaction take to execute? Faster Performance = Happier Customers Faster Performance = More Transactions HOWEVER Performance is the P in R-A-S-P and the other letters should be ignored to your peril! There is always a trade-off to consider!
- #12: LATENCY = Time delay in requesting an operation and it being initiated Key factor in large scale distributed applications Typically not taken into account during development