Problem

When creating a large number of Skupper v2 Listener and Connector resources (paired by routingKey), all CRs appear in the API, but it takes a long time until every resource reports Ready=True.

Reproduce

• Repo scenario: skupper/tests/e2e/scenarios/service-scale

• Command (example):

  make test TEST="service-scale" \
    EXTRA_VARS="-e service_count=300 -e ready_retries=400 -e ready_delay=3"

• Cluster: kind (single control plane), MetalLB, Skupper v2 controller (dev cluster setup).

Observed behavior

• Full playbook wall time on the order of ~15–17 minutes for 300 pairs; a large fraction is spent in Ansible waits:

  • Wait for all Connectors Ready (east)
  • Wait for all Listeners Ready (west)

• While investigating live with kubectl:

  • Connector objects often showed Pending / not Ready with No matching listeners until the paired Listener side was present and reflected in Skupper’s matching logic. Ready is not implied by “CR exists.”
  • After all CRs existed, Listener Ready counts increased gradually and sometimes stalled before completing, suggesting a reconciliation / status-propagation tail

From the API model, Ready for these resources depends on per-resource Configured work and Matched to the peer (HasMatchingListener / HasMatchingConnector) derived from the linked network view. Even when all CRs are already created, finishing Configured + updating matching + writing status for hundreds of resources appears to take many minutes.