Home/ Questions/Q 126345
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-11T05:18:11+00:00

I have a large ASP.NET website on a hosted platform. It shares the machine

  • 0

I have a large ASP.NET website on a hosted platform. It shares the machine with a lot of other applications. We do not have access to the machine itself (only an FTP account).

Our client is complaining that it is starting to perform rather badly, particularly around peak hours. I’ve run some remote measurements (using a JMeter-like tool) that tells me that, yes, it does indeed perform rather badly during peak hours. It doesn’t tell me why though. The client is resisting a move to a dedicated server without some hard facts.

As I see it, what I need are hard data about the machine itself. Setting up a local performance test environment would be extremely time-consuming, and I have no way to estimate the server performance.

My question: is there a good way to collect (a lot) of performance measurements when I have limited access to the machine, and certainly no access to the performance monitor? Any code would have to run in the asp.net application itself, without screwing it up too much.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. We had a similar problem with our asp.net application hosted on a shared server, which also started to perform badly during peak hours.

    Although I don’t know of an elegant solution to your question, this is what we did:

    1. Talk to your host providers to see what additional information they can give you – it’s in their best interest to keep their clients happy. Our host providers were able to give us some time with one of their network engineers who provided us with some decent CPU and memory utilization stats.

    2. Take your own performance measurements by dumping information to either a log file (using log4net) and/or the database – for example, user sessions, search times, page hits, timing measurements around key functionality. From this information we were able to ascertain what our systems normal behavior was for a set number of automation tests.

    3. Setup a local server (not necessarily same stats as hosted/production server) with your application loaded and give it a full load/performance/capacity testing (we used Red Gate’s ANTS Profiler). The stats that you gather from that will give you and your client a good indication of how the system should behave under certain loads with a known environment. Yes, this can be time consuming but it will give you a great performance measuring tool so that you can catch/fix bottlenecks locally rather than on production.

    Good luck.

    • 0