-
Notifications
You must be signed in to change notification settings - Fork 228
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bring back --enable-exec for declarative function pipelines. #2302
Comments
Can you share more details about the overheads that container runtime introduces ? Regarding performance, yes, there is scope for improvements here and we will be addressing it soon. An unrelated question: What tooling/SDK did you (or your team) use to develop the functions ? |
You have to pull each container and you need a container run time. Would you rather have one container with the Docker daemon mounted that then pulls and executes 20 containers, or would you rather have one container with 20 binaries packaged in it that you can run directly? Try a simple example on your own terminal. A pipeline with 10 functions ran via exec can finish in seconds, 10 containers that aren't already pulled takes much longer.
Performance is fine for
We used |
Describe your problem
We use Kpt internally, where we have nearly 20 functions. We produce all of the function binaries that we use and we put them into our pipeline images so that Kpt can use them directly.
A typical invocation might look like
kpt fn run --enable-exec $DIR/
for us.We explicitly do not want to use the container runtime because it introduces lots of overhead we don't have to deal with, as well as the performance of the pipeline falling off a cliff relative to the exec runtime.
The decision to neuter this functionality in Kpt v1 was a mistake that ignores enterprise users and favors toy demos / public pipelines.
The text was updated successfully, but these errors were encountered: