LECTURE 5
MONITORING WORKSTATION
PERFORMANCE
1
2
Involved Topics
Monitoring Server Memory
Monitoring Processor Usage
Monitoring Disks
Monitoring Network Usage
3
Monitoring Server Memory
Why Monitor Server Memory?
How to Identify and Resolve Memory Bottlenecks
How to Monitor Memory
4
The Important of Monitoring the
Server Memory
Monitor server memory to determine
whether any of the following conditions exist:
Memory bottlenecks
Insufficient memory
Excessive paging
Memory leaks
5
How to Identify and Resolve Memory
Bottlenecks?
Acceptable
Memory counter Desired value Action
average range
Find the process that is causing
Pages/sec 0–20 Low paging
Add RAM
Minimum of 5% of total Find the process that is using RAM
Available Bytes High
memory Add RAM
Find the process that is using RAM
Committed Bytes Less than physical RAM Low
Add RAM
Remain steady, Check for memory leak
Pool Nonpaged Bytes Not applicable
no increase in application
Find the process that is causing
Page Faults/sec Below 5 Low paging
Add RAM
6
Monitoring Processor Usage
What Is Processor Usage?
How to Identify and Resolve Processor Bottlenecks
How to Monitor Processor Usage
7
What Is Processor Usage?
Percentage of time that the processor is working
Monitor to detect processor bottlenecks
Tool Counter Display
Task Manager CPU Usage Graph
Percentage of elapsed
Performance %Processor Time
time to run non-idle thread
8
How to Identify and Resolve Processor
Bottlenecks
Acceptable
Processor counter Desired value Action
average range
Find process using excessive
% Processor Time Less than 85% Low processor time
Upgrade or add another processor
System: Processor
Less than 10 Low Upgrade or add additional processor
Queue Length
Find process using excessive
Server Work Queues:
Less than four Low processor time
Queue Length Upgrade or add another processor
Find controller card generating
Interrupts/sec Depends on processor Low
interrupts
9
Monitoring Disks
Why Monitor Disks?
How to Identify and Resolve Disk Bottlenecks
How to Monitor Disks
10
Why Monitor Disks?
Monitor disks to determine:
Presence of disk bottlenecks
Need for disk defragmentation
Need for additional or faster disks
Presence of excessive paging
Disk efficiency
11
How to Identify and Resolve Disk
Bottlenecks
Acceptable Desired high or
Physical disk counter Action
average range low value
Monitor to see if paging is
% Disk Time Under 50% Low occurring
Upgrade disk subsystem
Current Disk Queue
0–2 Low Upgrade disk subsystem
Length
Avg. Disk Bytes/Transfer Baseline or higher High Upgrade disk subsystem
Disk Bytes/sec Baseline or higher High Upgrade disk subsystem
12
Monitoring Network Usage
What Is Network Usage?
How to Identify and Resolve Network Bottlenecks
How to Monitor Network Usage
13
What Is Network Usage?
Percentage of network bandwidth in use on the
segment being monitored
Monitoring network usage helps you detect network
bottlenecks
Bottlenecks in network communications directly affect
the experience of the user at the client workstation and
the entire network
Typical causes of network bottlenecks are:
◦ Overloaded server
◦ Overloaded network
◦ Loss of network integrity
14
How to Identify and Resolve Network
Bottlenecks
Acceptable Desire high or low
Network interface counter Action
average range value
Network Utilization (in Task Generally lower than
Low Low
Manager) 30%
Network Interface: Upgrade network adapter or
Baseline or higher High
Bytes Sent/sec physical network
Perform further analysis to
Network Interface: Bytes
Baseline or higher High determine cause of problem
Total/sec
Upgrade or add another adapter.
Less than 50% of the
Server: Bytes Upgrade network adapter or
capacity of the bandwidth NA
Received/Sec physical network
of the network card
15
Guidelines for Using Counters and
Thresholds
Subsystem Counter Threshold
Monitor page faults Over 5 per second
Monitor available RAM Less than 5% of total
Monitor committed bytes More than physical RAM
% Processor time, % Privileged Time, % User
Above 85%
Time
Above 2
System: Processor Queue Length
Above 2
Server Work Queues: Queue Length
If more than 50%, check for
% Disk Time
excessive paging
Current Disk Queue Length
Greater than 2
Server: Bytes Total/sec, Network Interface: Higher than the baseline
Bytes Total/sec number
16log on to:
For detail,
http://technet.microsoft.com/en-us/library/cc759363(WS.10).aspx
Setting Monitoring Configuration
17
18
Best Practices for Monitoring Server
Performance
Set up Performance Logs and Alert
Keep monitoring overhead low
Analyze performance results and establish a performance baseline
Set alerts
Perform system tuning
Monitor trends for capacity planning and add or upgrade
components
19
Example of Server Monitoring Tools
Scalyr’s Server Monitoring:
to monitor your logs for subtle or impending problems, whether your business runs
on a single server or is distributed across hundreds of servers
intelligent alerting, and log analysis tools let you detect and resolve attacks
quickly to minimize risk and downtime
optimized for lightning-fast search and analysis, across all servers
20
How Scalyr Works?
