You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, resources for PrivateLoadZone test runs are set only during creation of PLZ as part of its spec. To change them, one has to delete and re-create PrivateLoadZone object. This is a cumbersome way of dealing with resources and not well suited for the PLZ as a long-term CRD.
Suggested Solution (optional)
The simplest way to improve this approach is to make resources mutable: all PLZ test runs that are started after the edit of PLZ object are to use new values of resources.
A more complex option is to switch to a different way of resource definition in PLZ -- with a resource pool. In general, the test runs can be different: some of them are heavy and require a lot of resources, while other test runs are very small, e.g. with VU in {1, 10}, and could be executed with less resources. Having a resource pool available for PLZ test runs, GCk6 can then decide what kind of CPU and memory to designate for a specific test run.
Switching to a resource pool would imply having a more complex logic in GCk6 but might result in a more efficient usage of computational resources overall.
This second option, with the pool, is currently more of an idea rather than a set course of action, and feedback on topic would be welcome.
Already existing or connected issues / PRs (optional)
Feature Description
Currently,
resources
for PrivateLoadZone test runs are set only during creation of PLZ as part of its spec. To change them, one has to delete and re-create PrivateLoadZone object. This is a cumbersome way of dealing with resources and not well suited for the PLZ as a long-term CRD.Suggested Solution (optional)
The simplest way to improve this approach is to make
resources
mutable: all PLZ test runs that are started after the edit of PLZ object are to use new values of resources.A more complex option is to switch to a different way of resource definition in PLZ -- with a resource pool. In general, the test runs can be different: some of them are heavy and require a lot of resources, while other test runs are very small, e.g. with VU in
{1, 10}
, and could be executed with less resources. Having a resource pool available for PLZ test runs, GCk6 can then decide what kind of CPU and memory to designate for a specific test run.Switching to a resource pool would imply having a more complex logic in GCk6 but might result in a more efficient usage of computational resources overall.
This second option, with the pool, is currently more of an idea rather than a set course of action, and feedback on topic would be welcome.
Already existing or connected issues / PRs (optional)
The mutability of PLZ resources came up here:
The text was updated successfully, but these errors were encountered: