Is dotTrace suitable to use on Production Servers?
We have a performance issue with a web application that is only occurring on Production servers. Is dotTrace suitable for profiling in production environments or will it impact performance and not advisable to use there?
If there are any articles or discussions about using dotTrace in production environments, please point to them.
Please sign in to leave a comment.
It will slowdown an app under profiling of course, but not dramatically. "Timeline" profiling mode is the most lightweight but it depends on the number of events produced by the app. "Sampling" profiling mode is more intrusive but its overhead is more even distributed and predictable. Anyway the overhead should not exceed 10%.
As far as I understand you are experiencing slowdown of your app, so slowing it down a little more at this particular time is not a very big deal. Thus I would recommend to use a command line version of dotTrace, and attach it to your app at exactly the moment when problems occurs, for several seconds. Usually it's enough to identify a problem.
Out of curiosity, could you tell a couple of words how you determine that your app has a performance issue? Is it an increasing of the response time or you see it by some monitoring tool?
Hi Ed,
thank you for that advice. Its an increase in response time of visual indicators that we see.
There is a Web UI where users can upload files for processing and add metadata about the files, then they click submit when ready for processing. There are a couple of visual indicators: one in the web UI tells the user when processing is complete. But also, processed files are supposed to appear in another system but we can see that is also delayed. The submitted files are normally processed within a few minutes but a certain type of file can sometimes take 30-40 minutes to process which should not happen, and this only occurs in production.
And it is erratic: sometimes the delay will happen and other times there is no issue.
Thanks for the details.
If you can determine this situation automatically, you can run console dotTrace by the script, as I said, at exact moment of the problem and be sure that your production will not stop for a significant time.