Highload Documentation

[vokhuongcse0604]

Ask your question here:

I have a question about the "High Scale" section in the document Request Flow.
I am confused about the equation "current_demand + target-burst-capacity > (pods * concurrency-target)" being used to represent traffic dropping. I believe the correct equation should be "current_demand < (pods * concurrency-target) + target-burst-capacity" because the incoming traffic (current_demand) must be lower than the setting capacity to route traffic to activator pods. Could you please explain if I am misunderstanding this case?

[skonto]

Hi @vokhuongcse0604,

Target burst capacity is a global and per-revision integer setting that determines the size of traffic burst a Knative application can handle without buffering

So that definition means until you have that capacity from the actual application pods, Activator will do the buffering for you (activator will be on the path, proxy mode). If you do have enough pods, you transition to serve mode or high scale.
Please take a look at the related blog post https://knative.dev/blog/articles/demystifying-activator-on-path.

[github-actions]

This issue is stale because it has been open for 90 days with no
activity. It will automatically close after 30 more days of
inactivity. Reopen the issue with /reopen. Mark the issue as
fresh by adding the comment /remove-lifecycle stale.