Restrict unneccessary calculation of unused threshold/summary metrics #764
Labels
Comments
na--
added a commit
that referenced
this issue
Oct 12, 2018
This, when used in conjunction with --no-thresholds, should reduce the k6 memory usage for long-running tests. This is something like an all-or-nothing approach to #764 - even though it's not the best solution, it's pretty easy to implement and support and we can always extend it later.
na--
added a commit
that referenced
this issue
Oct 16, 2018
This, when used in conjunction with --no-thresholds, should reduce the k6 memory usage for long-running tests. This is something like an all-or-nothing approach to #764 - even though it's not the best solution, it's pretty easy to implement and support and we can always extend it later.
|
Closing this in favor of #1321 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Sometimes it may be a waste time and memory to calculate the summary stats for some or all of the metrics. For example, if there are no percentile-based thresholds for a metric and the user doesn't care about the end-of-test summary (because they use an external output like InfluxDB or Load Impact Cloud or they use the LI cloud execution). Adding an option that allows users to specify whether they want to calculate those stats could lighten the CPU and memory load in those situations.
Depending on the performance and memory characteristics of HdrHistogram, implementing this may end up being unnecessary, but I decided to add it as an issue so we consider it as well. And it's somewhat of a lightweight version of the full metric filtering, though there are use cases where you might want one but not the other.
The text was updated successfully, but these errors were encountered: