eValid -- Automated Web Quality Solution
Browser-Based, Client-Side, Functional Testing & Validation,
Load & Performance Tuning, Page Timing, Website Analysis,
and Rich Internet Application Monitoring.
eValid -- Monitoring Mode Usage Recommendations
This page gives some basic recommendations
about using eValid in monitoring mode.
should allow eValid to operate trouble-free 24x7.
- Use The Correct Operating System
Here is a table of support status for
Windows OS Versions.
- Use The Windows Scheduler
Launch eValid's from Windows batch files that are run
directly from the Windows scheduler.
It's easy to set the starting time and the repeat interval.
eValid launches done from the Windows scheduler are very reliable.
- Use Batch Files For Multiple Launches
If you want to run multiple eValid copies
all starting at the same time,
do this from a Windows batch file
that is launched by the scheduler (as described above).
You can use the START eValid -B script.evs ...
command based on eValid's
Command Line Interface.
The Windows command line interpreter uses
the most reliable eValid playback process-start method.
- Use Small Delays to Space Out Multiple Launches
We have found that many errors occur when a batch script launches
too many eValid copies at one time.
The solution in most cases is to provide a minimal delay between each
launch using a command like SLEEP -m 250,
which sleeps 250 msecs,
between each START command in your batch launch script.
Take care to avoid operating system conflicts.
- Avoid File Name Conflicts
When two or more eValid copies are competing for access to the same-named script file
or log file then may create a Windows "file sharing violation".
Make sure parallel eValids use different filenames.
- Organize Tests into Different Folders
It is good practice,
and also simplifies maintenance of your scripts,
if you declare a separate Project folder for all of your eValid scripts,
and then have a separate Group folder for each individual
type of test you are running, so that your actual script.evs and the
associated logfiles are found at ".../Project/Group/..files.."
Neatness pays off.
A neat ship is a happy ship!
- Provide Enough RAM
Use as the base RAM figure the
Available Physical Memory value from the
Windows Task Manager's Performance tab.
Allow 10-20 MB RAM for each eValid instance
that you are planning on running in parallel.
For example, if you want to have 10 eValid's execute in parallel,
make sure you have at least 150 MB RAM available.
This will assure all eValid copies execute in memory;
swapping should be avoided.
until it is unable to do more work.
- To Calculate The Maximum Monitoring Capacity
To calculate the maximum number of tracks,
use Windows: System Tools > System Information
to the amount of Available Physical Memory.
This is the available RAM in total minus the RAM
that is used for Windows system functions.
Given that eValid will take 10-20 MB per copy,
keep the total number of parallel eValid copies
such that the total eValid memory usage is
BELOW 75% of the Available Physical Memory.
A 50% margin is even better.
- Consider Use Of The -RT Switch
Consider use of the eValid -RT switch to kill off any eValid executions
that don't finish playback in the specified time.
Note that the -RT switch will NOT affect eValid copies that experience
system loading errors or issues that
produce Modal popup messages that have to be dismissed by hand.
Schedule your machine to keep
too many multiple (parallel) eValid's from causing
- Beware Of Cascading Failures
If you see "Unable to Open Document" modal popups, they are probably
due to insufficient resources for the Windows loader to load eValid.
You have to dismiss such modals manually.
(eValid can suppress popups but at that point, eValid is not running).
Once you get one or two of these messages,
you're nearly guaranteed to eventually run out of resources
and clog the machine with more and more processes.
- Consider Using eValidP.exe
The eValidP.exe version of eValid -- eValidP.exe --
is an eValid engine intended for playback only.
It has a ~15% smaller runtime footprint,
and runs ~20% faster than the full eValid.
You don't have to worry:
eValidP.exe behaves 100% identically to eValid.
- Turn off Detailed Timings
The playback "detailed timings" EventLog feature is known in some cases that involve redirections of parts of
a page (e.g. when images are being served from alternative domains) to require excessive time
and or to cause a type of browser lockup that only is broken when the browser is brought into
and/or out of focus.
Normally the detailed timing data is not needed in monitoring
and it is therefore best to turn this capability OFF by selecting the Standard or Simple level.
- Switch Your Machine to Use The "Classic Look"
If you are on XP or Vista the fancy graphics do consume extra resources.
Switch to the "Windows Classic Style" by right-clicking on your screen and changing the settings.
This often makes a +20% improvement.
- Change the Machine Optimization Mode
From the Control Panel, using the System option, you can choose the Advanced Options tab that
"maximize for performance."
We have seen up to 20% improvement with this option.
- Consider Using eValid's LoadTest Mode CPU Operation Changes
If you are attempting to launch a large number of
eValid's in parallel,
you may need to apply the
Machine Adjustments for eValid LoadTest Operation.
We have seen up to 20% improvement with this option.
- What Are Realistic Limits?
As you can see from the above,
realistic limits on how many eValid playbacks you can obtain
per hour involve all of the following factors and considerations:
- the power and capacity of your computer;
- the speed of your internet connection;
- the available physical memory that can be used by eValid copies;
- the total number of eValid copies running in parallel;
- the length of each eValid playback (RAM usage grows with time);
- the amount of data you're collecting (in eValid logfiles);
- the "safety factor" of available extra CPU time.
It is MUCH better to have a machine with some extra capacity than to have it clog up due
to lack of resources.
- Installation Location Issues
We have heard of some strange behavior due to installation details.
Here is our recommendation:
Please do NOT use the Program~1 trick, but instead
use "Program Files" in your batch invocation.
We have heard that the ~1 may expend to a wrong (prior?) eValid.
Put eValid on your PATH, so there is no ambiguity about which
eValid you are running and which evalid.lic it is using.
It is a good idea to NOT put your scripts and logfiles
in the installation directory -- even though eValid will
work fine from there.
It is best to isolate playbacks so that there is no possibility
of two or more eValid's inadvertantly interfering with each other.
- Unusual Things To Watch Out For
Here are some unusual things to watch out for:
- If you see a slow growth in playback time in your script
as it runs over and over for serveral days, you may wish to
set the "days to keep pages in history" on your machine to zero.
The slowdown is due to the fact that if the list is very, very
long the browser has to spin through a [growing] list to update
the history records.
- Additional Support Pages
Please see the following support pages for additional information:
eValid cannot be held responsible for the outcome of monitoring
applications of eValid that do not conform to
recommended operating procedures.