Skip to content

APF: Understand the failures of 5k scalability tests with client-side rate-limitting disabled #115293

Open
@wojtek-t

Description

@wojtek-t

The results can be found in:
#109614

Facts:

  • latency metrics include waiting time in APF
  • so basically, long waiting time (saturated PLs) would result in exactly this

Hypothesis:

  • some PLs are saturated and thus latency metrics are exceeding the threshold

But we need to confirm this and figure out how to improve it.

[The next step would be to see how we can speed up this test even further and see what will happen later, but I think we're currently mostly blocked on scheduler throughput.]

/assign @MikeSpreitzer
/cc @tkashem

Metadata

Metadata

Assignees

Labels

needs-triageIndicates an issue or PR lacks a `triage/foo` label and requires one.sig/api-machineryCategorizes an issue or PR as relevant to SIG API Machinery.sig/scalabilityCategorizes an issue or PR as relevant to SIG Scalability.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions