Software Testing

Essential Performance Counters for Application Based on Windows Platform

2011-04-12T14:57:00.000+05:30

Essential Dot NET Performance Counter

2011-04-12T14:45:00.001+05:30

For list of others essential performance counters for applications based on Windows platform see here.

When there is a requirement of collecting performance counters during performance testing, it becomes tough to select few counters from existing many. All counters are important in its context and cannot be simply ignored. So, it is very important to recheck the list of counters that needs to be monitored before every performance run. You might need to change the counters as per yours requirements.

Following are some of the Dot NET counters, that I find bare essential to start with though. The list of counters and its preferred / threshold values are being collected from multiple relevant blogs / links and have been mentioned in the ‘References’ section below.

Performance Object	Performance Counter	Preferred Value
Exception	# of Exceps Thrown / Sec	< 5% of RPS (Requests / sec)
Loading	Rate of assemblies	None
Lock & Thread	Contention Rate / Sec	0 or very low
Lock & Thread	Current Queue Length	None
Memory	# Gen 0 Collections	10 * # Gen 1 Collections
	# Gen 1 Collections	1/10^th of Gen 0 Collections
	# Gen 2 Collections	1/10^th of Gen 1 Collections
	# Induced GC	As low as possible
	# of Pinned Objects	< 100
	% Time in GC	< 5%
	Gen 0 Promoted Bytes/Sec	As low as possible
	Gen 1 Promoted Bytes/Sec	As low as possible
	Large Object Heap size	As low as possible

Exception

Describe the performance counters that provide information about the exceptions thrown by an application.

# of Exceps Thrown / Sec

It displays the number of exceptions thrown per second. It includes both .NET exceptions and unmanaged exceptions that are converted into .NET exceptions.

Preferred Value: < 5% of RPS (Requests / sec)

Threshold Value: > 100

Notes:

Correlate this metric with ASP.NET applications \ Requests per second. Don't use _Global_ counter and use specific application though. Graph can be plotted against Elapsed Time vs. (Requests / sec, # Exceps Thrown / sec) and can be figured out if exceptions are occurring because of number of requests increasing or while running for a longer duration or exception happening from the start. If exceptions rate is higher from start then it might indicates programmer are throwing exception as part of normal processing.
The total number of managed exceptions thrown per second. As this number increases, performance degrades.
Exceptions should not be thrown as part of normal processing.
Response.Redirect, Server.Transfer, and Response.End all cause a ThreadAbortException to be thrown multiple times, and a site that relies heavily upon these methods will incur a performance penalty.

Loading

1. Describes the performance counters that provide information about assemblies, classes, and application domains that are loaded.

2. The _Global_ counter instance should not be used with these performance counters, because it is updated by all managed processes. Instead, use the aspnet_wp instance.

Rate of Assemblies

It displays the number of assemblies loaded per second across all application domains

Preferred Value: None

Threshold Value: None

Notes:

Correlate 'Rate of Assemblies' with 'Memory Availability'. If memory is being used because of increasing loading of assemblies then it can be recommended to load less number of user controls in an aspx page.

If the assembly is loaded as domain-neutral from multiple application domains, this counter is incremented only once.

Lock & Thread

Describes the performance counters provide information about managed locks and threads that an application uses.

Contention Rate / Sec

Displays the rate at which threads in the runtime attempt to acquire a managed lock unsuccessfully

Preferred Value: 0 or very low

Threshold Value: See preferred value

Notes:

Unsuccessful locks can cause serious performance issues when the rate is high, as the threads are not only synchronized but ultimately unsuccessful, potentially throwing exceptions and waiting excessively.

Current Queue Length

Displays the total number of threads that are currently waiting to acquire a managed lock in the application

Preferred Value: None

Threshold Value: None

Notes:

Presence of queue in a system definitely gives an hint of waiting. But no waiting also means the resources are not being utilized properly, so queue should always be correlated with utilization too.

Memory

Describes the performance counters provide information about the garbage collector

# Gen 0 Collections

Displays the number of times the generation 0 objects (that is, the youngest, most recently allocated objects) are garbage collected since the application started.

Preferred Value: 10 times more than #Gen 1 collections

Threshold Value: See preferred value

Notes:

The number of times generation 0 objects have been garbage collected. Objects that survive are promoted to generation 1.

A collection is performed when room is needed to allocate an object, or when someone forces a collection by calling System.GC.Collect. Collections that involve higher generations take longer, since they are preceded by collections of lower generations.

The # Gen N Collections counters and the % Time in GC counter are the best for identifying performance issues caused by excessive allocations.

# Gen 1 Collections

Displays the number of times the generation 1 objects are garbage collected since the application started.

Preferred Value: See threshold value

Threshold Value: 1/10^th of # Gen 0 collections

Notes:

Objects that survive are promoted to generation 2.

# Gen 2 Collections

Displays the number of times the generation 2 objects are garbage collected since the application started.

Preferred Value: See threshold value

Threshold Value: 1/10^th of # Gen 1 collections

Notes:

Generation 2 is the highest, thus objects that survive collection remain in generation 2.

Gen 2 collections can be very expensive, especially if the size of the Gen 2 heap is excessive.

# Induced GC

Displays the peak number of times garbage collection was performed because of an explicit call to GC.Collect.

Preferred Value: As low as possible

Threshold Value: See preferred value

Notes:

It is good practice to let the garbage collector tune the frequency of its collections

If # induced GC is increasing and indicates a high number then it indicates a problem. It is never a good idea to induce GC explicitly rather it is should be found out that why heap size is growing.

# of Pinned Objects

Displays the number of pinned objects encountered in the last garbage collection.

Preferred Value: Less than 100

Threshold Value: See preferred value

Notes:

A pinned object is one that the garbage collector cannot move in memory.

This counter tracks the pinned objects only in the heaps that are garbage collected.

If number of pinned objects are large in number (more than 100) or constantly increasing, it indicates problems related to programming and incoming memory issue

% Time in GC

Displays the percentage of elapsed time that was spent performing a garbage collection since the last garbage collection cycle

Preferred Value: Less than 5%

Threshold Value: An average of 5% or more; short-lived spikes larger than this are common.

Notes:

All threads are suspended during a garbage collection.

More GC collections means less real work

The most common cause of a high % Time in GC is making too many allocations on a per request basis.

The second most common cause is inserting a large amount of data into the ASP.NET cache, removing it, regenerating it, and reinserting it into the cache every few minutes.

Avoid allocating short-lived objects larger than 85,000 bytes. These objects are allocated in the large object heap, and are more expensive to collect.

Gen 0 Promoted Bytes/Sec

Displays the bytes per second that are promoted from generation 0 to generation 1

Preferred Value: As low as possible

Threshold Value: See preferred value

Notes:

Memory is promoted when it survives a garbage collection.

This counter is an indicator of relatively long-lived objects being created per second.

If this value is large than it means lots of long-lived objects are being created per second and it needs to be looked into it whether they can be disposed early or not.

Gen 1 Promoted Bytes/Sec

Displays the bytes per second that are promoted from generation 1 to generation 2

Preferred Value: As low as possible

Threshold Value: See preferred value

Notes:

Objects that are promoted only because they are waiting to be finalized are not included in this counter.

This counter is an indicator of very long-lived objects being created per second.

Nothing is promoted from generation 2 because it is the oldest generation.

Large Object Heap size

Displays the current size, in bytes, of the Large Object Heap

Preferred Value: As low as possible

Threshold Value: See preferred value

Notes:

Objects greater than 20 KB are treated as large objects by the garbage collector and are directly allocated in a special heap; they are not promoted through the generations.
If this value is increasing, it indicates lots of BLOBs are getting creating and which needs to be looked into it.

References:

1. Performance Counters in .NET Framework

2. ASP.NET Performance Monitoring and when to Alert Administrators

3. Investigating Memory Issues

4. Performance Counters Analyzing DNN

5. Identifying and Prevent Memory Leaks in Managed Code

6. Improving .NET Application Performance and Scalability

Essential WWW Service Performance Counter

2011-04-11T13:15:00.001+05:30

For list of others essential performance counters for applications based on Windows platform see here.

Following are some of the WWW Service counters, that I find bare essential to start with though. The list of counters and its preferred / threshold values are being collected from multiple relevant blogs / links and have been mentioned in the ‘References’ section below.

Performance Object	Performance Counter	Preferred Value
Bytes (Sent, Received, and Transferred) Counters	Bytes Sent / sec	None
Bytes (Sent, Received, and Transferred) Counters	Bytes Total/sec	None
Connections and Attempts Counters	Current Connections	As high as possible
Connections and Attempts Counters	Connection Attempts/sec	None
Requests Counters	Total Method Requests/sec	As high as possible
CGI Requests and ISAPI Extension Requests Counters	ISAPI Extension Requests/sec	None
File Counters for the Web Service Cache Object	File Cache Hits %	As high as possible
	File Cache Misses	As low as possible
	File Cache Flushes	As low as possible
URI Counters for the Web Service Cache Object	URI Cache Misses	As low as possible
	URI Cache Hits %	None
	URI Cache Flushes	As low as possible
Kernel Counters for the Web Service Cache Object	Kernel: URI Cache Hits%	As high as possible
	Kernel URI Cache Misses	As low as possible
	Kernel: URI Cache Flushes	As low as possible

Bytes (Sent, Received, and Transferred) Counters

Bytes Sent / sec

The rate, in seconds, at which data bytes have been sent by the WWW service

Preferred Value: None

Threshold Value: None

Notes:

It indicates the throughput of web service.
If throughput is not increasing or getting plateau on increasing the number of total methods requests / sec, then it means, WWW service is not able to handle all http requests and few requests are getting rejects / queued
Data collected by the IIS 6.0 service counters underestimates the total number of bytes that the IIS 6.0 services actually transmit to the network. Because these values are collected at the application layer, they measure data only. They do not measure protocol headers, control packets, or retransmitted bytes.
In general, the bytes counted by the services represent 60 to 70 percent of the total number of bytes transmitted by the services on the network. If the sum of bytes for all services accounts for two-thirds or more of total network bandwidth, you can assume that your network is running at or near the total capacity of its communications link

Bytes Total/sec

The sum of Bytes Sent/sec and Bytes Received/sec

Preferred Value: None

Threshold Value: None

Notes:

Correlate with NIC bandwidth and find out if there is any network related bottleneck because of WWW service.

The value of it should be as high as possible.

Connections and Attempts Counters

Current Connections

The number of active connections to the WWW service

Preferred Value: As high as possible

Threshold Value: None

Notes:

Each connection that an IIS 6.0 service establishes consumes processor time. The network adapter interrupts the processor to signal that a client has requested a connection. Further processing is required to establish and maintain the connection, to fulfill client requests sent over the connection, and — when the connection is closed — to delete the structures that serviced the connection. Each time a connection is established, the load on the server increases. A curve in the graph of the Current Connections counter on any IIS 6.0 service performance objects that reaches a high value and then plateaus. This pattern often indicates that additional connections are being blocked or rejected.

Ideally when number of concurrent users increase, current connections should also follow. In case, current connections are not increasing or getting plateau, which might means maximum connection limit has been reached and rest of the users connections are being rejected.

Connection Attempts/sec

The rate, in seconds, at which connections to the WWW service have been attempted since the service started

Preferred Value: None

Threshold Value: None

Notes:

In case of failure point, it suggests the connections that were being attempted and provide threshold information in that term (connection attempts / sec). It can be important metrics as it can be associated with performance requirements also.

It can also provide a basis for validating the load on to the system

Requests Counters

Total Method Requests/sec

The rate, in seconds, at which all HTTP requests have been received

Preferred Value: As high as possible

Threshold Value: None

Notes:

It can be interesting to compare this with "ASP.NET Applications\Requests/sec" and "Web Service\ISAPI Extension Requests/sec" in order to see the percentage of static pages served versus pages rendered by aspnet_isapi.dll.

If numbers of current connections are increasing but ‘Total Methods Requests / sec’ is not increasing than it means there are connections which are idle / waiting and not invoking any further requests. In that case http keep-alive settings need to be looked into it. In a way, it represents idle sessions.

CGI Requests and ISAPI Extension Requests Counters

ISAPI Extension Requests/sec

The rate, in seconds, at which ISAPI extension requests are being processed by the WWW service

Preferred Value: None

Threshold Value: None

Notes:

Used primarily as a metric for diagnosing performance issues. It can be interesting to compare this with "ASP.NET Applications\Requests/sec" and "Web Service\Total Method Requests/sec." Note that this includes requests to all ISAPI extensions, not just aspnet_isapi.dll.

When number of total methods requests / sec increases, ISAPI extension requests should also increase. If not then application server seems to be queuing these requests.

File Counters for the Web Service Cache Object

The IIS file cache stores file contents in memory. If the file is larger than the 256 KB default maximum file size, IIS creates a file handle, passes the handle to HTTP.sys, and then closes the handle. The maximum file size is controlled by the MaxCachedFileSize registry key.

File Cache Hits %

The ratio of user-mode file cache hits to total cache requests that have been made since the WWW service started up.

Preferred Value: As high as possible

Threshold Value: None

Notes:

This value might be low if the Kernel URI cache hits percentage is high.

It provides IIS user-mode cache utilization, if the utilization is low it should be correlated with File Cache Misses also as it provides the real picture

If File Cache Hits % are low but kernel level URI Cache Hits% are more then it is good thing but if both are low then it means IIS caching is not being used properly

File Cache Misses

The numbers of unsuccessful lookups in the user-mode file cache that has been made since the WWW service started.

Preferred Value: As low as possible

Threshold Value: None

Notes:

Misses is always counted whether kernel mode is used or user mode cache is used

File Cache Flushes

The number of files that have been removed from the user-mode cache since the WWW service started

Preferred Value: As low as possible

Threshold Value: None

Notes:

If flushes are more then IIS user-mode cache might not be utilized properly

URI Counters for the Web Service Cache Object

The user-mode URI cache is known as the URI cache. The URI cache stores metadata about a URL. Metadata can include a variety of configuration data, including data about headers and authentication. This configuration data is stored in the IIS Admin service. The URI cache stores the metadata as an object. For each request that reaches IIS user mode, the worker process checks the URI cache to see if the object for the requested URL has been cached. If the object has not been cached, the worker process must call the IIS Admin service for the appropriate metadata. This call is expensive from a performance and scalability standpoint. If the URI cache has a stored object for the URL, the request pulls in the object and continues processing

URI Cache Misses

The number of unsuccessful lookups that have been made in the user-mode URI cache since the WWW service started

Preferred Value: As low as possible

Threshold Value: None

Notes:

The value of this counter increases when IIS cannot locate requested content in either the HTTP.sys response cache or the IIS file cache, and must locate the requested content on the hard disk.

If the value of this counter is low, responses are being sent from either the IIS file cache or the HTTP.sys response cache.

URI Cache Hits %

The ratio of URI Cache Hits to total cache requests that have occurred since the WWW service started.

Preferred Value: None

Threshold Value: None

Notes:

Depends on content. If content cannot be cached in the HTTP.sys response cache, the value of this counter should be as high as possible. If your content can be stored in the HTTP.sys response cache, the value of this counter should be low.

Cache hits percentage should be more but at the same time correlated it with kernel side caching hits as well

URI Cache Flushes

The total number of URI cache flushes that have occurred since the WWW service started

Preferred Value: As low as possible

Threshold Value: None

Notes:

More flushes mean cache is not being utilized properly.

Kernel Counters for the Web Service Cache Object

1. HTTP response cache is known as kernel URI cache.

2. Only responses that will be cached in the kernel are responses to HTTP GET requests.

3. ISAPI responses can also be cached in kernel by setting some response headers.

Kernel: URI Cache Hits%

The ratio of Kernel: URI Cache Hits to total cache requests since the WWW service started.

Preferred Value: As high as possible

Threshold Value: None

Notes:

Applies to static unauthenticated content and dynamic content that is marked as cacheable.

Hits percentage tells about cache utilization

Kernel URI Cache Misses

The number of unsuccessful lookups in the kernel URI cache that have occurred since the WWW service started.

Preferred Value: As low as possible, relative to the number of requests

Threshold Value: None

Notes:

Each request for dynamic content increases the value of the counter by one

Kernel: URI Cache Flushes

The number of kernel URI cache flushes that have occurred since the WWW service started

Preferred Value: As low as possible, relative to the number of requests

Threshold Value: None

Notes:

This number increases every time a file is flushed from the HTTP.sys response cache (HTTP.sys is the kernel-mode device driver in IIS 6.0), which means that the content has not been accessed in the past 2–4 minutes. The only way to decrease this number is to flush the cache less often, although frequent flushing can cause HTTP.sys to use more memory for content that is not being accessed.

References:

1. Monitoring ASP.NET Application (IIS 6.0)

2. ASP.NET Performance Monitoring, and When to Alert Administrators

4. Monitoring Connections (IIS 6.0)

5. Working With http.sys or Kernel Mode Caching in Internet Information Services 6-0

6. URI Cache

7. File Cache

Essential ASP.NET Performance Counter

2011-04-09T18:33:00.001+05:30

For list of others essential performance counters for applications based on Windows platform see here.

Following are some of the ASP.NET counters, that I find bare essential to start with though. These counters are related to specific web applications and not system wide, which have been ignored intentionally. The list of counters and its preferred / threshold values are being collected from multiple relevant blogs / links and have been mentioned in the ‘References’ section below.

Performance Object	Performance Counter	Preferred Value
Cache API Counters	Cache API Hit Ratio	> 80%
Cache API Counters	Cache API Turnover Rate	None
Errors Counters	Errors during Execution	None
Errors Counters	Errors Unhandled During Execution/sec	None
Output Cache Counters	Output Cache Hit Ratio	None
Output Cache Counters	Output Cache Turnover Rate	None
Requests Counters	Requests in Application Queue	None
Requests Counters	Requests / sec	None
Miscellaneous Counters for ASP.NET Applications	Pipeline Instance Count	None

Cache API Counters

Cache API Hit Ratio

It indicates ratio of cache hits to misses of objects called from user code.

Preferred Value: Greater than 80%

Threshold Value: See preferred value

Notes:

A low ratio can indicate inefficient use of caching techniques in the program.
This counter does not track use of the cache internally by ASP.NET
With sufficient RAM, you should normally record a high (greater than 80 percent) cache hit ratio.

Cache API Turnover Rate

The number of additions to and removals from the cache per second, when it is used through the external APIs (excluding internal use by the ASP.NET framework)

Preferred Value: None

Threshold Value: None

Notes:

A high turnover rate indicates that items are being quickly added and removed, which can be expensive.
High turnover suggest inefficient utilization of cache. Cache API hit ratio can be greater than 90% but if turnover rate is high then it's an indication of inefficient utilization of cache.

Errors Counters

Errors during Execution

The total number of errors that occurred during the execution of an HTTP request

Preferred Value: None

Threshold Value: None

Notes:

It excludes parser and compilation errors though.
It indicates failure during run-time and should be monitored.
Correlate it with ‘Errors Unhandled during Execution / sec’ to find out if application is handling negative flows with the help of catching / throwing exceptions.

Errors Unhandled During Execution/sec

The total number of unhandled exceptions per second at run time

Preferred Value: None

Threshold Value: None

Notes:

Unhandled exceptions are more problematic in terms of performance.

Output Cache Counters

Output Cache Hit Ratio

The percentage of total requests that were serviced from the output cache

Preferred Value: None

Threshold Value: None

Notes:

The total hit-to-miss ratio of Output Cache requests.

Output Cache Turnover Rate

It indicates the number of additions and removals to the output cache per second

Preferred Value: None

Threshold Value: None

Notes:

A high turnover rate indicates that items are being quickly added and removed, which can be expensive.

Requests Counters

Requests in Application Queue

The number of requests in the application request queue

Preferred Value: None

Threshold Value: None

Notes:

In addition to Requests Current, Requests in Application Queue provides a warning for when requests will be rejected.
If there are only a couple virtual directories, increasing the default appRequestQueueLimit to 200 or 300 may be suitable, especially for slow applications under heavy load.
Presence of queue most of the time suggests performance issues, so if queue is not decreasing or increasing further investigation is required.

Requests/sec

The average number of requests that have been executed per second

Preferred Value: None

Threshold Value: None

Notes:

This counter represents the current throughput of the application.
Correlate this counter with number of requests from clients and find out if it is decreasing / increasing accordingly. If not, then it needs to be analyzed.

Miscellaneous Counters for ASP.NET Applications

Pipeline Instance Count

The number of active request pipeline instances for the specified ASP.NET application. It represents the maximum number of concurrent requests that are being processed for a specific application.

Preferred Value: None

Threshold Value: None

Notes:

Only one execution thread can run within a pipeline instance, this number represents the maximum number of concurrent requests that are being processed for a specific application.
In most circumstances, it is better for this number be low when the server is busy, because this means that the CPU is well used.
The number of pipeline instances should be steady. Sudden increases are indicative of back end latency.

References:

1. IIS

2. ASP.NET Performance Monitoring, and When to Alert Administrators

3. Performance Counters for ASP.NET

4. How ASP.NET Works

5. Measuring .NET Application Performance

Essential MS SQL Server Performance Counter

2011-04-06T15:19:00.000+05:30

Following are some of the MS SQL Server counters, that I find bare essential to start with though. The list of counters and its preferred / threshold values are being collected from multiple relevant blogs / links and have been mentioned in the ‘References’ section below.

Performance Object	Performance Counter	Preferred Value
Access Methods	Forwarded records / sec	< 10 out of 100 batch requests / sec
	Full scans / sec	(Index Searches / sec) / (Full scans / sec) > 1000
	Index searches / sec	(Index Searches / sec) / (Full scans / sec) > 1000
	Page splits / sec	< 20 per 100 Batch requests / sec
	Table lock escalations / sec	None
	Workfiles created / sec	None
	Worktables created / sec	< 20
Buffer Manager	Buffer cache hit ratio	> 90%
	Checkpoint pages / sec	< 300
	Free list stalls / sec	< 2
	Free pages	> 640
	Lazy writes / sec	< 20
	Page life expectancy	> 300
	Page lookups / sec	(Page lookups / sec) / (Batch requests / sec) < 100
	Page reads / sec	< 90
	Page writes / sec	< 80
Databases	Log growths	None
Databases	Transactions / sec	None
General Statistics	Logins / sec	< 2
	Logouts / sec	< 2
	User connections	None
Latches	Average latch wait time (ms)	< 1
	Latches waits / sec	(Total Latch Wait Time) / (Latch Waits / Sec) < 10
	Total latch wait time (ms)	(Total Latch Wait Time) / (Latch Waits / Sec) < 10
Locks	Average wait time (ms)	< 500
	Lock requests / sec	< 1000
	Lock timeouts (timeouts > 0) / sec	Nil
	Lock wait time (ms)	None
	Lock waits / sec	Nil
	Number of deadlocks / sec	Nil
Cache Manager	Cache hit ratio	> 90%
SQL Errors	Errors / sec	Nil
SQL Statistics	Batch requests / sec	< 1000
	SQL compilations / sec	< 10% of batch requests / sec
	SQL recompilations / sec	< 10% of SQL compilations / sec

Access Methods

Forwarded records / sec

It indicates number of records fetched through forwarded record pointers.

Preferred Value: Less than 10 out of 100 batch requests / sec

Threshold Value: See preferred value

Notes:

High value of this counter can be avoided by using default values, using char instead of varchar i.e. variable size string.
Tables with NO clustered index can fetch records through forwarded record pointers. If you start out with a short row, and update the row creating a wider row, the row may no longer fit on the data page. A pointer will be put in its place and the row will be forwarded to another page.
Adding clustered index is simplest way to eliminate problem
Occur when row / record moved from one database page to another because changed record cannot fit back in original page

Full scans / sec

It indicates the number of unrestricted full scans. These can either be base table or full index scans.

Preferred Value: (Index Searches / sec) / (Full scans / sec) > 1000

Threshold Value: See preferred value

Notes:

It indicates Unrestricted linear searches through table or index for example "SELECT * FROM TABLE"
SQL Query "Estimated Execution Plan" can identify them ahead of time.
SQL Query "Actual Execution Plan" and SQL Profiler (trace) can identify them when they occur
Compare Full scans / sec with Forwarded records / sec
Performance can be reduced because of too high disk access.
If we see high CPU then we need to investigate this counter, otherwise if the full scans are on small tables we can ignore this counter.
Few of the main causes of high full scans / sec are Missing indexes, Too many rows requested

Index searches / sec

It indicates number of index searches.

Preferred Value: (Index Searches / sec) / (Full scans / sec) > 1000

Threshold Value: See preferred value

Notes:

Index searches are used to start range scans, single index record fetches, and to reposition within an index.

Page splits / sec

It indicates number of page splits per second that occur as a result of overflowing index pages.

Preferred Value: < 20 per 100 Batch requests / sec

Threshold Value: See preferred value

Notes:

It can be avoided by having proper fill factor.
Interesting counter that can lead us to our table / index design. This value needs to be low as possible
If you find that there is a very strong correlation between page splits and excessive disk I/O activity, then most likely you have a page splitting problem.
The problem can be remedied by Proper fill factor, Rebuild yours indexes more often, Get a faster I/O subsystem

Table lock escalations / sec

It indicates the number of times locks on a table were escalated.

Preferred Value: None

Threshold Value: None

Notes:

Row, key or page locks automatically escalated to coarser table locks as appropriate
A high number needs revisit to the query and the indexes on the table

Workfiles created / sec

It indicates number of work files created per second.

Preferred Value: None

Threshold Value: None

Notes:

Work files could be used to store temporary results for hash joins and hash aggregates.
It helps in identifying performance issues related to tempDB growth

Worktables created / sec

It indicates number of work tables created per second.

Preferred Value: Less than 20

Threshold Value: 20 or more

Notes:

Work tables could be used to store temporary results for query spool, LOB variables, XML variables, and cursors.
It can be used to monitor tempDB growth.
Temporary tables created in tempdb by SQL Server to perform a sort, union or group by operation.

Buffer Manager

Buffer cache hit ratio

It indicates the percentage of pages that were found in the buffer pool without having to incur a read from disk.

Preferred Value: Greater than 90%

Threshold Value: 85% or less

Notes:

When this percentage is high, your server is operating at optimal disk I/O efficiency. If this value decreases over time, you might consider adding physical memory to the server.
If the DBA is seeing low readings for the buffer cache hit ratio, the Page Life Expectancy statistic should be checked.

Checkpoint pages / sec

It indicates number of pages flushed by checkpoint or other operations that require all dirty pages to be flushed.

Preferred Value: < 300

Threshold Value: 300 or more

Notes:

When a checkpoint occurs, all dirty pages are written to disk. This is a normal procedure and will cause this counter to rise during the checkpoint process. What you don't want to see is a high value for this counter over time. This can indicate that the checkpoint process is running more often than it should, which can use up valuable server resources. If this has a high figure (and this will vary from server to server), consider adding more RAM to reduce how often the checkpoint occurs, or consider increasing the "recovery interval" SQL Server configuration setting.
This value is relative, it varies from server to server, we need to compare the average to a base line capture to tell if the value is high or low.

Free list stalls / sec

It indicates the number of requests that had to wait for a free page.

Preferred Value: < 2

Threshold Value: 3 or more

Notes:

It is the frequency with which requests for available database pages are suspended because no buffers are available.

Free pages

It indicates total number of pages on all free lists.

Preferred Value: > 640

Threshold Value: Consistently close to zero

Notes:

It is an Indicator of insufficient SQL server memory
If Free Pages shows an average number that is rather poor, but none of the other counters that tracks memory show a bad number, then you’re a-okay.

Lazy writes / sec

It indicates number of buffers written by buffer manager's lazy writer.

Preferred Value: < 20

Threshold Value: 20 or more

Notes:

This counter tracks how many times a second that the Lazy Writer process is moving dirty pages from the buffer to disk in order to free up buffer space. Generally speaking, this should not be a high value, say more than 20 per second or so. Ideally, it should be close to zero. If it is zero, this indicates that your SQL Server's buffer cache is plenty big and SQL Server doesn't have to free up dirty pages, instead waiting for this to occur during regular checkpoints. If this value is high, then a need for more memory is indicated

Page life expectancy

It indicates number of seconds a page will stay in the buffer pool without references.

Preferred Value: Greater than 300

Threshold Value: 300 or less

Notes:

If the value is less it indicates SQL cache is cold, Memory problems, Missing indexes etc.

Page lookups / sec

It measures number of times database attempted to find a page in buffer pool.

Preferred Value: (Page lookups / sec) / (Batch requests / sec) < 100

Threshold Value: See Preferred Value

Notes:

It performs a logical read rather than the physical read.
Useful for corroborating and further quantifying buffer cache hit ratio.
Compare page reads / sec with page lookups / sec.
When the value much greater than preferred value then it is an indication that while query plans are looking up data in the buffer pool, these plans are insufficient. Identify queries with the highest amount of logical I/Os and tune them.

Page reads / sec

It indicates number of physical database page reads issued.

Preferred Value: Less than 90

Threshold Value: See preferred value

Notes:

It measures physical I/O and not logical I/O.
High value of this counter might indicates Insufficient database memory, Application improperly accessing database, Improper database table implementation or indexing or memory constraints

Page writes / sec

It indicates number of physical database page writes issued.

Preferred Value: Less than 80

Threshold Value: 90 or more

Notes:

80 – 90 per second is normal, anything more we need to check the lazy writer/sec and checkpoint counters, if these counters are also relatively high then, it’s memory constraint.

Databases

Log growths

It indicates total number of log growths for this database.

Preferred Value: None

Threshold Value: None

Notes:

If there is lot of activity in this counter we need to allocate static and large enough space for our log files.

Transactions / sec

It indicates the number of transactions started for the database

Preferred Value: None

Threshold Value: None

Notes:

It indicates the average utilization of the server

General Statistics

Logins / sec

It indicates total number of logins started per second.

Preferred Value: Less than 2

Threshold Value: 2 or more

Notes:

Greater than 2 per second indicates that the application is not correctly using 'Connection Pooling'

Logouts / sec

It indicates total number of logouts started per second.

Preferred Value: Less than 2

Threshold Value: 2 or more

Notes:

Greater than 2 per second indicates that the application is not correctly using 'Connection Pooling'

User Connections

It indicates the number of users connected to the system.

Preferred Value: None

Threshold Value: None

Notes:

It is recommended to review this counter with "Batch Requests / sec". A surge in "user connections" may result in surge of "Batch Requests / sec"
With a blocking problem, for example you might see user connections, lock waits and lock wait time all increase while batch requests per second decreases.

Latches

Average latch wait time (ms)

It indicates average latch wait time (milliseconds) for latch requests that had to wait.

Preferred Value: Less than 1

Threshold Value: 1 or more

Notes:

It indicate large number of physical I/Os or long I/O times
If value greater than 1 then often coincide with low buffer cache hit ratio
If this number is high, yours server might have resource limitations.

Latches waits / sec

It indicates number of latch requests that could not be granted immediately and had to wait before being granted.

Preferred Value: (Total Latch Wait Time) / (Latch Waits / Sec) < 10

Threshold Value: See Preferred Value

Notes:

None

Total latch wait time (ms)

It indicates total latch wait time (milliseconds) for latch requests that had to wait in the last second.

Preferred Value: (Total Latch Wait Time) / (Latch Waits / Sec) < 10

Threshold Value: See Preferred Value

Notes:

None

Locks

Average wait time (ms)

It indicates the average amount of wait time (milliseconds) for each lock request that resulted in a wait.

Preferred Value: Less than 500

Threshold Value: 500 or more

Notes:

Waiting hints for performance issues, lower the value the better it is.
If the value goes higher then 500, there may be blocking going on; we need to run blocker script to identify blocking.
A high value indicates Memory Pressure, IO Subsystem problem, Improper Indexes, Improper SQL Structures, Improper placement of files in the disk subsystem, Incorrect usage of SQL isolation level

Lock requests / sec

It indicates the number of new locks and lock conversions requested from the lock manager

Preferred Value: Less than 1000

Threshold Value: 1000 or more

Notes:

It indicates the number of requests for a type of lock per second. Lock requests/sec > 1000 indicates that the queries are accessing large number of rows; the next step is to review high read queries. If you also see high Avg. Wait time, then it’s an indication of blocking, then review the blocking script output.

Lock timeouts (timeouts > 0) / sec

It indicates number of lock requests that timed out. This does not include requests for NOWAIT locks.

Preferred Value: Nil

Threshold Value: None

Notes:

It indicates # of lock requests that exceed maximum specified wait time.

Lock wait time (ms)

It indicates total wait time (milliseconds) for locks in the last second

Preferred Value: None

Threshold Value: None

Notes:

For “Lock Wait Time” it is recommended to look beyond the Avg value. Look for any peaks that are close (or exceeds) to a wait of 60 sec.
Though this counter counts how many total milliseconds SQL Server is waiting on locks during the last second, but the counter actually records at the end of locking event. So most probably the peaks represent one huge locking event. If those events exceeds more than 60seconds then they may have extended blocking and could be an issue. In such cases, thoroughly analyze the blocking script output. Some applications are written for timing out after 60 seconds and that’s not acceptable response for those applications.

Lock waits / sec

It indicates number of lock requests that could not be satisfied immediately and required the caller to wait before being granted the lock.

Preferred Value: 0

Threshold Value: More than 0

Notes:

This counter reports how many times users waited to acquire a lock over the past second.
If this value is nonzero then it is an indication that there is at least some level of blocking occurring. If you combine this with the Lock Wait Time counter, you can get some idea of how long the blocking lasted.
Correlate this counter with Lock wait time (ms)

Number of deadlocks / sec

It indicates number of lock requests that resulted in a deadlock

Preferred Value: 0

Threshold Value: Greater than zero

Notes:

SQL profiler can provide information about how deadlock was created.

Cache Manager

Cache Hit Ratio

It indicates ratio between cache hits and lookups

Preferred Value: Greater than 90%

Threshold Value: Less than 85%

Notes:

Lower values indicate too many ad-hoc queries
Lower values often associated with high values of SQL compilations / sec and SQL re-compilations / sec

SQL Errors

Errors / sec

It indicates number of errors per second.

Preferred Value: Nil

Threshold Value: None

Notes:

Monitoring errors always provide quality aspects of application.

SQL Statistics

Batch requests / sec

It indicates the number of SQL batch requests received by server

Preferred Value: Less than 1000

Threshold Value: 1000 or more

Notes:

This counter is required for other correlation (e.g. SQL compilations / sec).
Batch requests / sec can be compared with System Context Switches / sec to highlight need for SQL server connection affinity
Generally over 1000 batch requests / sec, indicates a very busy SQL server
From a network bottleneck approach, a typical 100Mbps network card is only able to handle about 3000 batch requests / sec. If you have a server that is this busy, you may need to have 2 or more network cards.
Sometimes low batch requests/sec can be misleading. If there were a SQL statements/sec counter, this would be a more accurate measure of the amount of SQL Server activity. For example, an application may call only a few stored procedures yet each stored procedure does lot of work. In that case, we will see a low number for batch requests/sec but each stored procedure (one batch) will execute many SQL statements that drive CPU and other resources.

SQL compilations / sec

It indicates number of SQL compilations

Preferred Value: Less than 10% of total Batch Requests / sec

Threshold Value: 40% of total Batch Requests / sec

Notes:

If stored procedure requested after removal or invalidation, it is recompiled.
High compilation rates frequently corresponds with lower Cache Manager: Cache Hit Ratios
High compilation rates indicates lack of stored procedure usage
High compilation rates indicates possible memory shortage
If you see a high value such as over 100, then it’s an indication that there are lots of adhoc queries that are running, might cause CPU usage, solution is to re-write these adhoc as stored procedure or use sp_executeSQL

SQL Re-compilations / sec

It indicates number of SQL re-compilations

Preferred Value: Nil

Threshold Value: 10 % or more of SQL compilations / sec

Notes:

A recompile can cause deadlocks and compile locks that are not compatible with any locking type.

References

1. Performance Monitoring - Basic Counters

2. Performance Counters

3. Top SQL Server 2005 Performance Counters Issues

4. Performance Counters

5. How to perform a SQL Server Performance Audit

6. Database Monitoring Samples

7. SQL Server - Performance Counters Guidance

8. Forum

Essential Windows Performance Counters

2011-03-31T15:09:00.002+05:30

Following are some of the windows counters, that I find bare essential to start with though. The list of counters and its preferred / threshold values are being collected from multiple relevant blogs / links and have been mentioned in the ‘References’ section below.

Performance Object	Performance Counter	Preferred Value
Logical Disk	% Free Space	> 10%
Memory	Available Bytes	> 20% of Physical RAM
	% Committed Bytes in Use	< 80%
	Page Reads / sec	< 5
	Pages Writes / sec	As low as possible
Physical Disk	% Disk Time	< 80%
	Avg. Disk Queue Length	None
	Current Disk Queue Length	(Queue Length - # Spindles on the disks) < 3
	Disk Bytes / sec	As high as possible
	Disk Transfers / sec	< 50 per physical disk
	Split IO / sec	None
Process	% Processor Time	< 70%
	Private Bytes	< 60% of RAM
	Working Set	None
	Virtual Bytes	None
Processor	% Interrupt Time	< 50%
	% Privileged Time	None
	% Processor Time	< 80%
	Interrupts/sec	< 1000
System	Context Switches/sec	1500 – 3000 per processor
	Processor Queue Length	< 3 * (Number of Processors)
	System Calls/sec	None
Network Interface	Bytes Total / sec	< 50% of Current Bandwidth of NIC

Logical Disk

% Free Space

It reports the percentage of unallocated disk space to the total usable space on the logical volume.

Preferred Value: > 10%

Threshold Value: < 5%

Notes:

It is advisable to monitor all logical volumes than _Total instance. For example, C drive gets filled by some default log files or historical data and slows down the overall system.

There is no % Free Space counter for the Physical Disk object, so in case it is required to collect and monitor ‘Free Space’ of physical disk than logical volumes than _Total instance can be used.

Memory

Available Bytes

It shows the amount of physical memory (in bytes) immediately available for allocation to processes or for system use.

Preferred Value: > 20% of physical RAM

Threshold Value: < 4 MB for sustained period of time

Notes:

As the available counter decreases, paging increases.

If the Committed Bytes don’t decrease and Available Bytes don’t increase, when the program is closed then the system may have memory leak.

% Committed Bytes in Use

It shows the ratio of Committed Bytes to Commit Limit.

Preferred Value: < 80%

Threshold Value: < 90%

Notes:

If Available Bytes are decreasing and % Committed Bytes in Use is increasing and doesn’t return back to initial state then it might suggests memory leaks in a program.

Page Reads / sec

It shows the rate, in incidents per second, at which the disk was read to resolve hard page faults.

Preferred Value: Less than 5

Threshold Value: Greater than 5 for sustained period of time

Notes:

It is the primary indicator of memory shortage.

It shows the number of read operations, without regards the number of pages retrieved in each operation. For finding number of pages reads per operation, monitor Pages Input / sec as well.

Pages Writes / sec

Similar to Page Reads/sec, this counter indicates how many times the disk was written to in an effort to clear unused items out of memory.

Preferred Value: As low as possible

Threshold Value: None

Notes:

Increasing values in this counter often indicate a building tension in the battle for memory resources.

It should always be reported with Page Reads / sec to corroborate the finding of memory constraint.

Pages are written to disk only if they are changed while in physical memory, so they are likely to hold data and not code.

Physical Disk

% Disk Time

It indicates the percentage of time the hard disk is busy.

Preferred Value: Less than 80%

Threshold Value: 80%

Notes:

It indicates overall disk utilization.

An average value of 80% or above indicates that the hard disk can’t keep up with the demand. It could be due to a hard disk that is too slow, or it could be caused by excessive paging.

If it’s busy almost all the time, and there is a large queue, the disk might be a bottleneck.

Avg. Disk Queue Length

This counter displays %Disk Time as a decimal with no defined maximum. (A %Disk Time of 100% equals an Avg. Disk Queue Length of 1.0.)

Preferred Value: None

Threshold Value: None

Notes:

This counter will be needed in times when the disk configuration employs multiple controllers for multiple physical disks. In these cases, the overall performance of the disk I/O system, which consists of two controllers, could exceed that of an individual disk. Thus, if you were looking at the %Disk Time counter, you would only see a value of 100%, which wouldn't represent the total potential of the entire system, but only that it had reached the potential of a single disk on a single controller. The real value may be 120% which the Avg. Disk Queue Length counter would display as 1.2.

Current Disk Queue Length

It is an indication of the number of transactions that are waiting to be processed.

Preferred Value: (Queue Length - # Spindles on the disks) should be less than 3

Threshold Value: (Queue Length - # Spindles on the disks) should be greater than 3 for sustained period of time

Notes:

Even if % Disk Time is 100% but if Current Disk Queue Length is less than it means disk is being utilized properly.

The queue is an important measure for services that operate on a transaction basis. Just like the line at the supermarket, the queue is representative of not only the number of transactions, but also the length and frequency of each transaction.

Disk Bytes / sec

It indicates the rate at which bytes are transferred and is the primary measure of disk throughput.

Preferred Value: As high as possible

Threshold Value: None

Notes:

It is the primary measure of disk throughput.
It answers the question, how fast is data being moved (in bytes)?

Disk Transfers / sec

It measures the number of read and writes completed per second, regardless of how much data they involve.

Preferred Value: Less than 50 per physical disk

Threshold Value: 50 per physical disk

Notes:

It answers the question, how fast are data requests being serviced?
It measures disk utilization

Split IO / sec

It shows the rate, in incidents per second, at which input/output (I/O) requests to the disk were split into multiple requests.

Preferred Value: None

Threshold Value: None

Notes:

A split I/O might result from requesting data in a size that is too large to fit into a single I/O, or from a fragmented disk subsystem.
On single-disk systems, a high rate for this counter tends to indicate disk fragmentation.

Process

% Processor Time

It provides processor utilization information by a process.

Preferred Value: Less than 70%

Threshold Value: 70%

Notes:

If one process is utilizing more than 70% of processor then it means its not giving any time to other processes.

Private Bytes

It shows the size, in bytes, that this process has allocated that cannot be shared with other processes.

Preferred Value: Less than 60% of RAM

Threshold Value: 60% of RAM

Notes:

If a memory leak is occurring, this value will tend to steadily rise.

Working Set

The working set is the set of memory pages that were touched recently by the threads in the process.

Preferred Value: None

Threshold Value: None

Notes:

If free memory in the computer is above a threshold, pages are left in the working set of a process, even if they are not in use.
When free memory falls below a threshold, pages are trimmed from working sets. If the pages are needed, they will be soft-faulted back into the working set before leaving main memory.
The size of the working set will grow and shrink as the VMM can permit. When memory is becoming scarce the working sets of the applications will be trimmed. When memory is plentiful the working sets are allowed to grow.
Larger working sets mean more code and data in memory making the overall performance of the applications increase. However, a large working set that does not shrink appropriately is usually an indication of a memory leak.

Virtual Bytes

It shows the size, in bytes, of the virtual address space that the process is using.

Preferred Value: None

Threshold Value: None

Notes:

If the program with the memory leak is allocating virtual memory in its own address space, the memory leak should be evident by tracking per process Virtual Bytes Counter. If the amount of Virtual Bytes allocated for a process increases steadily over the life of a process, there is good reason to suspect a leak.

Processor

% Interrupt Time

It shows the percentage of time that the processor spent receiving and servicing hardware interrupts during the sample interval.

Preferred Value: Less than 50 %

Threshold Value: 50 %

Notes:

If this value exceeds 50% of the processor time then there might be some hardware issues present.

% Privileged Time

It is the amount of time the processor was busy with Kernel mode operations.

Preferred Value: None

Threshold Value: None

Notes:

If the processor is very busy and this mode is high, it is usually an indication of some type of NT service having difficulty

% Processor Time

It shows the percentage of elapsed time that this thread used the processor to execute instructions.

Preferred Value: Less than 80 %

Threshold Value: 80 %

Notes:

It provides processor utilization information.
Correlate it with System \ Processor Queue Length to make sure that when utilization is more than 80% then queue is also getting developed otherwise there is no issue with processor utilization.

Interrupts/sec

It shows the numbers of interrupts the processor was asked to respond to.

Preferred Value: Less than 1000

Threshold Value: Greater than 1000 for sustained period of time

Notes:

A sustained value over 1000 is usually an indication of a problem. Problems would include a poorly configured drivers, errors in drivers, excessive utilization of a device (like a NIC on an IIS server), or hardware failure.
Compare this value with the System: Systems Calls/sec. If the Interrupts/sec is much larger over a sustained period, then probably there is a hardware issue present in the system.

System

Context Switches/sec

It indicates the rate at which the processors switch thread contexts.

Preferred Value: 1500 – 3000 per processor

Threshold Value: Greater than 6000 per processor

Notes:

A high number may indicate high lock contention or transitions between user and kernel mode.
A high context-switch rate often indicates that there are too many threads competing for the processors on the system.
Context Switches/sec should increase linearly with throughput, load, and the number of CPUs. If it increases exponentially, there is a problem.
Context switches occur when a running thread voluntarily relinquishes the processor, or is preempted by a higher priority, ready thread.

Processor Queue Length

It indicates how many threads are ready, but have to wait for a processor.

Preferred Value: Less than 3 * (Number of Processors)

Threshold Value: Greater than 3 * (Number of Processors) for sustained period of time

Notes:

There is a single queue for processor time, even on computers with multiple processors. Therefore, if a computer has multiple processors, you need to divide this value by the number of processors servicing the workload.
Unlike disk queue counters, it counts only waiting threads, not those being serviced.
A sustained processor queue of greater than two threads generally indicates processor congestion.

System Calls/sec

It measures the number of calls made to the system components, Kernel mode services.

Preferred Value: None

Threshold Value: None

Notes:

Indirectly it measure, how busy the system is taking care of applications and services—software stuff.

When compared to the Interrupts/Sec it will give you an indication of whether processor issues are hardware or software related.

Network Interface

Bytes Total/sec

It indicates the rate at which bytes are sent and received over each network adapter, including framing characters.

Preferred Value: Less than 50% of Current Bandwidth of NIC

Threshold Value: 50% of Current Bandwidth of NIC

Notes:

It tells overall how much information is going in and out of the interface.
On identification of NIC bottleneck, Bytes Received / sec & Bytes Sent / sec can be tracked later.

References:

Windows Server 2003 Performance Testing

How to find memory leak using performance monitor?

Detecting memory leak

Detecting processor bottleneck

Monitoring context switches

Detecting disk bottlenecks

Getting Started

Network Health

Keep tabs on yours network traffic

Performance monitor counters

Working with disk counters

Monitoring activities on multiple processor system

Windows page file and SQL server

ASP.NET performance monitoring

Key performance counters and their thresholds

Suggested performance counters to watch

Detecting cache bottleneck

RAM, virtual memory and page file, all that stuff

Always Specify a Content Type

2011-02-09T18:55:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

o Before browser can begin to check for a character set, they must first find the content type of the document being processed. If it is not specified, browser tries to identify using various algorithms, which adds an overhead of extra processing time

o To avoid any security vulnerability

How

o Configure web server to specify the content type in the HTTP headers

o If configuring of web server is not possible, then set the content type in the meta tag and specify the markup in the section of the HTML document

Exceptions

None

Recommendations

None

Specify Unique Character Encoding

2011-02-09T18:50:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

Why

o Specifying a character set differently in HTTP header and HTML markup, confuses the browser and incurs additional delays in redrawing.

How

Always set the same character set for both HTTP header and in the HTML markup

Exceptions

None

Recommendations

None

Specify a Character Set Early

2011-02-09T18:45:00.003+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

HTML documents are sent over the Internet as a sequence of bytes accompanied by character encoding information. Character encoding information is specified in the HTTP response headers or in the HTML markup of the document itself. The browser uses the character encoding information to convert the stream of bytes into characters that it renders on screen. Most browsers buffer a certain number of bytes before executing any JS or drawing. If they don’t find the character encoding information in the buffer they assume one by default and start rendering. Later if it doesn’t match, a browser reparses the input and redraws the page.

Why

o Specifying a character set early in the HTML document, allow browser to begin executing scripts immediately

How

o Configure web server to specify the character set in the HTTP headers

o If configuring of web server is not possible, then set the char set in the meta tag and specify the markup in the 'head' section of the HTML document

Exceptions

None

Recommendations

None

Specify Width and Height of All Images

2011-02-09T18:41:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

Specifying width and height of images in the web page eliminate the need to unnecessary reflows and repaints

How

Set the dimensions on the ‘img’ element itself or a block level parent. If the parent is not block-level, the dimensions will be ignored

Exceptions

None

Recommendations

None

Put CSS in the Document Head

2011-02-09T18:36:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

o By ensuring that style sheets are downloaded and parsed first, allows browser to progressively render the page

o Browser block rendering a web page until all external style sheets have been downloaded

o Inline style blocks can cause reflows and shifting of content, which adds an overhead of redrawing

How

o Put 'style' blocks in the 'head' section

o Always put external style sheets in the 'head' section using the 'link' tag

Exceptions

None

Recommendations

None

Avoid CSS Expressions

2011-02-09T17:55:00.000+05:30

For viewing all tips see here

Applicable To

IE version 5 – 7, which supports CSS expressions

How Improves Page Load Time

By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

CSS expressions allow dynamically changing document properties in response to various events. It consists of Java Script expressions embedded as the value of CSS properties in CSS declarations. Because of it, the browser reevaluates each CSS expression on trigger of every event. The performance degrades significantly if events are like ‘onMouseMove’, ‘window resize’ etc.

How

o Don’t use CSS expressions if possible

o Use JavaScript instead for dynamic styling

Exceptions

None

Recommendations

None

Use Efficient CSS Selectors

2011-02-09T17:50:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

o To speed up page rendering by minimizing the number of look ups

o To evaluate fewer rules by CSS rule engine

How

o Define rule which are specific as possible as the less specific the key, the greater the number of nodes that needs to be evaluated.

o Avoid rules with descended selectors as key as for each element that matches the key, the browser must also traverse up the DOM tree, evaluating every ancestor element until it finds a match or reaches the root tree

o Avoid rules with child or adjacent selectors as for each matching element, the browser has to evaluate another node

o Prefer class and ID selectors over tag selectors

o Use a class to apply a style to multiple elements

o Avoid redundant qualifiers like ID selectors qualified by class / tag selectors or class selectors qualified by tag selectors

o Use class selectors instead of descendent selectors

Exceptions

None

Recommendations

None

Serve Resources from a Consistent URL

2011-02-09T17:45:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

o By optimizing number of network roundtrips

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

o Sometimes it is necessary to reference same resource at multiple places in the web page. If the URL for these resources in not unique, multiple requests will be send by browser for the same resources. For this reason, resources should be served from consistent URL only.

o It eliminates duplicate downloads

o To remove additional DNS lookup time, if the resource is served from different hosts

How

Serve shared resources from a consistent URL across all pages in a site

Exceptions

None

Recommendations

None

Serve Only Properly Sized Images

2011-02-09T17:35:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing network latency (HTTP Payload Size)

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

Serving an image that is larger than the dimensions used in the markup instances, add unnecessary bytes to get downloaded

How

Use image editor for scaling the images as required and serve only those images when requested

Exceptions

If two different dimensions are being used in the markup for the same image then it makes sense to send only one image which is larger in size.

Recommendations

None

Optimize Images

2011-02-09T17:29:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

By optimizing network latency (HTTP Payload Size)

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

Images saved from some programs contain unnecessary comments and use too many colors and hence improperly optimized images take more space than they need to.

How

o By cropping unnecessary spaces

o By reducing color depth to the lowest acceptable level

o By removing comments in images

o By saving image to an appropriate format.

o By using GIF for very small or simple graphics and for images containing animations other use PNG

o By not using BMP or TIFFs

o By using image compressors

Exceptions

None

Recommendations

None

Defer Loading of JavaScript

2011-02-09T15:16:00.000+05:30

For viewing all tips see here

Applicable To

All Browsers

How Improves Page Load Time

o By optimizing parallelization of external resources

o By optimizing network latency (HTTP Payload Size)

o By optimizing browser rendering

Prologue

[Detail of the technology / concept being used in the tip if required]

Why

o Before browser can start rendering a web page, Java script code must be downloaded, parsed and executed first, which if it is not needed can be deferred

o Even when JS file is cached, browser blocks processing of all elements after the JS code, until browser load the code from disk and processed it

o For some browsers, while java scripts are being processed they block downloading all other external resources, which add considerable latency for script intensive application

How

o By allowing deferred bytes to be loaded asynchronously in the background

o By deferring all JS codes that are required for user – trigger events like onMouseClick

o By identifying all JS functions that are not actually used by the document before the onLoad event

o By forcing external JS file to be loaded after the onload event

Exceptions

None

Recommendations

Use DOM element for defer loading than usual simple HTML script to avoid cross-browser and same-domain policy issues