立即与支持人员聊天
与支持团队交流

Foglight Transaction Recorder 5.6.13 - Recorder User Guide

Overview of Foglight synthetic monitoring Use the Foglight Transaction Recorder Create and upload Foglight Transaction Recorder scripts View Foglight Transaction Recorder results Settings How Do I... Appendix: Best practices Appendix: Troubleshoot the Foglight Transaction Recorder Appendix: API for FTR Timer Usage

Gather network statistics when using Windows Vista or a later version

To gather network statistic from the Foglight Transaction Recorder when using Windows Vista or a later version, you must run as an elevated user or disable the User Account Control.

Scale Foglight Transaction Player Agents

The following topics provides procedures on how to scale Foglight Transaction Player Agents.

•
Maximum number of concurrent scripts
•
Determining maximum concurrent scripts
•
Assigning worker processes

Maximum number of concurrent scripts

The maximum number of concurrent scripts that can be run on a Foglight Transaction Player agent host is defined below. Exceeding these numbers can result in client-side degradation and response time measurements that are not consistent with (that are slower than) the actual performance of the monitored application. The number of CPU equals the number of CPU cores on the Foglight Transaction Player agent host machine (a dual-core processor counts as two CPUs).

Table 13. Maximum number of concurrent scripts
Number of CPUs
1

8

2

14

3

19

4

23

5

26

6

29

7

31

8

33

The numbers in the table above are general guidelines, not hard rules. CPU is generally the limiting factor when determining the maximum number of concurrent scripts, but memory consumption can also come into play. Memory and CPU usage on a Foglight Transaction Player agent host can be monitored with an OS cartridge or by examining data supplied in the Foglight Transaction Player intelligent agent. Longer scripts on complex web sites can use considerably more CPU and/or memory than short scripts monitoring simple web sites.

Each implementation is different so some customers may be able to run more scripts (and some fewer), concurrently than show in the above table. Faster CPUs are capable of running more scripts concurrently than slower CPUs. A more precise method for determining the maximum number of concurrent scripts is described below.

The maximum number of concurrent scripts running on a Foglight Transaction Player agent host is determined by the total worker processes assigned to the various agents on the host. Five Foglight Transaction Player agents each assigned two worker processes can never execute more than 10 Foglight Transaction Player scripts at any given time.

While there is no hard limit to the number of worker processes that can be assigned to a single Foglight Transaction Player agent, Quest recommends a maximum of five. If five worker processes is not sufficient to execute all of the scripts assigned to a Foglight Transaction Player agent, it is recommended that another agent be created and the scripts be split between them.

Likewise, there is no hard limit to the number of Foglight Transaction Player agents that can be configured on a single host. But since each agent uses as least one worker process, the practical limit is the same as the number of concurrent scripts/total worker processes.

Determining maximum concurrent scripts

The limiting factor on how many scripts can be running at any given time on a Foglight Transaction Player agent host is not determined by a certain percentage CPU or memory usage. The maximum number of concurrent Foglight Transaction Player scripts is exceeded when resource usage on the host machine begins to affect script response times. When this happens, slower response times are not due problem with the monitored system, but rather with the load put on the Foglight Transaction Player host machine. This is different for each implementation and can only be determined experimentally.

Two machines are required; the Foglight Transaction Player agent host to be tested, HostA, and a separate machine to act as a control, HostB. Record a few scripts with the Foglight Transaction Recorder that are representative of the scripts that will be used in production. Name them Script1, Script1, Script3. Configure Foglight Transaction Player agents, AgentA0 and AgentB0, on both machines to run a Script List, List0, consisting of Script1, Script2, and Script3. Start the agents. Let the scripts run for a few sample intervals. Response times for the scripts should be similar on both Foglight Transaction Player hosts.

The following provides procedures on determining the maximum number of concurrent scripts:
1
Stop all agents on both machines.
2
Create an new agent on HostA, name it Agent1.
3
Create a new Script List, List1 and assign three scripts - these can be the same or different script as used in List0. Start all agents on both machines and let them run for a few sample intervals. It is important that scripts are running concurrently, so be sure to start the agents simultaneously. The first run of the scripts can be slower than subsequent runs due to additional load on the host machine when the agents are restarted. Be sure to let the scripts several times before comparing response times. Response times on HostA, Agent0 should be similar to HostB, Agent0.
4
Repeat this process until HostA, AgentN shows response times that are significantly different that HostB, Agent0. At this point, HostA has exceeded its maximum number of concurrent scripts.
相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级