0

I'm trying to nail down why our application is sluggish at certain times. The application sometimes is zippy almost instantaneous and other times it is slow as molasses. The application is run on a server with clients in our department. It is only intranet communication. During the day after 10 am, it becomes super slow (until after 4 pm).

I am having the problem where the developers are pointing to our I.T. and our I.T. pointing fingers to our developers. Is there any way that a general user like me can isolate which is the major contributing factor? The only tool I am aware of using is the ping. Is there any other simple way to help give a clearer picture? That is the challenge, using tools that easily available on most computers.

1

A divide and conquer approach usually gives good results, you focus on a specific module (or a collection of modules) monitoring it or if possible replacing it with something else.

When you find a suspicious module try to further sub dividing it until you find the root cause.

Note that this is usually an exploratory and iterative approach, you try something that in return will hint what should be the next step.

The main downside of such an approach is in cases where the cause is not a single component but a chain of failures, or cases where you found the trigger to the problem but not the root cause.

| improve this answer | |
0

First of all you should to gather metrics like:

  • CPU consumption
  • RAM consumption
  • Disk I/O operations
  • Number of open connections

If your application works with database or with load balancer you should also gather metrics for the box hosting those components. For database case also track the number of open sessions.

All those values should be placed on a time scale so that you could match events basing on the timeline. This will give you some clue on what can cause issues.

P.S. - it also makes sense to analyze application logs for activities taking suspiciously long.

| improve this answer | |
0

I would over the course of a few days measure load times of the same pages at different times of the day. And simply provide that to the team and let them figure out what is going on.

For instance measure the same page or action at the same times everyday.

Action Time of Day Monday Tuesday Wednesday

Load Page A 9:00AM 5 sec. 4.5 sec. 5 sec.

Load Page A 1:00PM 30 sec. 45 sec. 35 sec.

Submit Record 8:45AM 10 sec. 12 sec. 9 sec.

Submit Record 2:30PM 45 sec. 60 sec. 45 sec.

Demonstrate your observations with data and let the team figure out what is causing it.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.