Any Performance Testers on board?

[SimonMac]

Our TA has been asked to look at how they are going to do the Performance Testing, Client wants it done in Prod but to me that's a RED FLAG, as the other environments are not representative due to being virtualised rather than physical tin I was wondering is there such a thing where a bench mark can be ran on Prod and the Virtual environment to give a comparison factor of performance, so when we run the Pref Testing on the non-prod environment we can use the same factor to give "comparable" results if it was run on Prod?

[stek]

No preprod env?

[SimonMac]

No preprod env?

Kinda, but in no fit state to be used.

[DimPrawn]

Are you trying to performance test the software on a fixed hardware scenario or performance test different hardware configurations with a fixed software release?

Are you looking at overall performance or are you load testing (scalability testing)?

If think you need to work out what you are trying to achieve....

[SimonMac]

Are you trying to performance test the software on a fixed hardware scenario or performance test different hardware configurations with a fixed software release?

Are you looking at overall performance or are you load testing (scalability testing)?

If think you need to work out what you are trying to achieve....

Me?! I'm trying to do sod all except get a supposed Technical Expert to stop asking me questions

[DimPrawn]

Me?! I'm trying to do sod all except get a supposed Technical Expert to stop asking me questions

Fight fire with fire.

Collate a big list of questions and fire them back.

Works for me.

[jmo21]

Fight fire with fire.

Collate a big list of questions and fire them back.

Works for me.

This!

[The Spartan]

Sounds really dodgey to me, I've done performance testing on a live environment out of hours. Normally we would have a pre-prod environment to run tests so we could do a comparison or benchmark it. These are often scaled down and through the magic of statistics we're able to deduce whether performance has been improved etc.

The problem with using the live environment is:
1) Are there any periods where customers don't use it
2) If you overload the system too much you could cause a lot of damage and possible downtime
3) You need to make sure the conditions are ideal before getting accurate results

[SimonMac]

Sounds really dodgey to me, I've done performance testing on a live environment out of hours. Normally we would have a pre-prod environment to run tests so we could do a comparison or benchmark it. These are often scaled down and through the magic of statistics we're able to deduce whether performance has been improved etc.

The problem with using the live environment is:
1) Are there any periods where customers don't use it
2) If you overload the system too much you could cause a lot of damage and possible downtime
3) You need to make sure the conditions are ideal before getting accurate results

They had no Environment Management for the first two Phases of a project, which is why I have been brought in for Phase 3, trying to play catch up getting all the existing Environments back in line