When using the HTTP samplers, both "follow redirects" and "download embedded resources" can generate multiple samples. These samples are presented as a parent sample with child samples which are used for the additional requests. The parent sample data is from the main response, but the parent sample size [1] and elapsed time are for the overall transaction.
In the case of "follow redirects", the child samples consist of all the request responses, including the final non-redirected response. However, in the case of "download embedded resources", only the embedded resources are added as child samples. There is no child sample that corresponds with the main page that was originally requested. The elapsed time and byte count for the main page are not directly available anywhere. [2] It seems to me that the sample results should really be consistent. I.e in both cases all responses should be shown as child samples, and the parent sample should reflect the overall size and elapsed time. The parent sample should still contain the data from the requested resource, otherwise Assertions and PostProcessors will break. Any views on this? Is it OK to change JMeter, or will that cause any problems? [1] At least that was the intention - as per Bug 45053 the initial response data bytes for "follow redirects" are counted twice. [2] The byte count can be derived by subtracting the other sample sizes, and the elapsed time can be estimated by looking at the start time of the first subsample. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
