CI build could fail with the below error right after the initialize stage.
Check if the delegate is installed in the same cluster where build is running, Try by adding delegate selector in infrastructure connector.. Delegate is not able to connect to the created build farm. Invalid request: No eligible delegate was able to confirm that it has the capability to execute [ x.x.x.x:20001]
During the CI pipeline execution, the delegate will create the build farm pod, and it will then try to connect to the lite-engine container (Internal Harness container) running within the same pod.
If the build farm is a GKE cluster with the intranode visibility configuration enabled, we might get into these issues as intranode visibility configures networking on each node in the cluster so that traffic sent from one Pod to another Pod is processed by the cluster’s Virtual Private Cloud (VPC) network, even if the Pods are on the same node. With this configuration enabled, you could have specific firewall rules blocking the communication between the build farm pod and the ephemeral build farm pod IP, which causes this issue.
More info about intranode visibility configuration in GKE cluster can be referred here
To isolate the issue, you could try disabling the intranode visibility option if you face this issue. If it works right after intranode visibility option is disabled, you can then apply necessary firewall rules to allow the delegate pods to communicate with the ephemeral build farm pod IPs.