《Systems Performance: Enterprise and the Cloud》读书笔记系列(九) —— 第二章(八)
2014-10-17 18:03
429 查看
1. Workload Characterization: workload characterization is a simple and effective method for identifying a class of issues: those
due to the load applied. It focuses on the input to the system, rather than the resulting performance. Your system may have no architectural or configuration issues present, but it is under more
load than it can reasonably handle.
2. Workloads can be characterized by answering the following questions:
1) Who is causing the load? Process ID, user ID? remote IP address?
2) Why is the load being called? Code path, stack trace?
3) What are the load characteristics? IOPS, throughput, direction(read/write), type? Include variance(stand derivation)
where appropriate.
4) How is the load changing over time? Is there a daily pattern?
It can be useful to check all of these, even when you have strong expectations about what the answer will be, because
you may be surprised.
3. The best performance wins are the result of eliminating unnecessary work.
4. Drill-Down Analysis: Drill-down analysis starts with examining an issue at a high level, then narrowing the focus
based on the previous findings, discarding areas that seem uninteresting, and digging deeper into those areas that are. A drill-down
analysis methodology has three stages:
1) Monitoring: This is used for continuously recording high-level statistics over time, and identifying or alerting
if a problem may be present.
2) Identification: Given a suspected problem, this narrows the investigation to particular resources or areas of
interest, identifying possible bottlenecks.
3) Analysis: Further examination of particular system areas is done to attempt to root-cause and quantify the issue.
5. Latency Analysis: Latency analysis examines the time taken to complete an operation, then breaks it into smaller
components, continuing to do subdivide components with the highest priority so that the root cause can be identified and qualified. Similar
to drill-down analysis, latency analysis may drill down through layers of the software stack to find the origin of latency issues.
due to the load applied. It focuses on the input to the system, rather than the resulting performance. Your system may have no architectural or configuration issues present, but it is under more
load than it can reasonably handle.
2. Workloads can be characterized by answering the following questions:
1) Who is causing the load? Process ID, user ID? remote IP address?
2) Why is the load being called? Code path, stack trace?
3) What are the load characteristics? IOPS, throughput, direction(read/write), type? Include variance(stand derivation)
where appropriate.
4) How is the load changing over time? Is there a daily pattern?
It can be useful to check all of these, even when you have strong expectations about what the answer will be, because
you may be surprised.
3. The best performance wins are the result of eliminating unnecessary work.
4. Drill-Down Analysis: Drill-down analysis starts with examining an issue at a high level, then narrowing the focus
based on the previous findings, discarding areas that seem uninteresting, and digging deeper into those areas that are. A drill-down
analysis methodology has three stages:
1) Monitoring: This is used for continuously recording high-level statistics over time, and identifying or alerting
if a problem may be present.
2) Identification: Given a suspected problem, this narrows the investigation to particular resources or areas of
interest, identifying possible bottlenecks.
3) Analysis: Further examination of particular system areas is done to attempt to root-cause and quantify the issue.
5. Latency Analysis: Latency analysis examines the time taken to complete an operation, then breaks it into smaller
components, continuing to do subdivide components with the highest priority so that the root cause can be identified and qualified. Similar
to drill-down analysis, latency analysis may drill down through layers of the software stack to find the origin of latency issues.
相关文章推荐
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(七) —— 第二章(六)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(五) —— 第二章(四)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(三) —— 第二章(二)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(二) —— 第二章(一)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(八) —— 第二章(七)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(六) —— 第二章(五)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(四) —— 第二章(三)
- 《Systems Performance: Enterprise and the Cloud》读书笔记系列(一) —— 第一章
- Systems Performance: Enterprise and the Cloud 读书笔记系列
- The C programming language --第二章 类型、运算符与表达式 读书笔记
- High Availability for the Ubuntu Enterprise Cloud (UEC) – Cloud Controller (CLC)
- Solaris 10 Advance Administrator 310-202 读书笔记 第二章 Describing the Client-Server Model(1)
- Drag-and-Drop into the Cloud
- Microsoft Enterprise Library 5.0 系列(八) Unity Dependency Injection and Interception
- 70-565 Pro: Designing and Developing Enterprise Applications Using the Microsoft .NET Framework 3.5 考试感言
- Autodesk云计算系列视频 --- 开篇介绍 Up to the cloud 直上云端
- Report from the Cloud Computing Summit: Business Intelligence, Data Integration, and the Cloud
- 70-565 Pro: Designing and Developing Enterprise Applications Using the Microsoft .NET Framework 3.5 考试感言
- Building Enterprise Applications with Windows Presentation Foundation and the
- Python Programming with the Java(TM) Class Libraries: A Tutorial for Building Web and Enterprise App