AWS VPC peering is created for VPCs in single region by referring to aws docs.

The diagram below explains the same.

Both the VPC peering connections are active and their route tables adjusted for subnets. But when tried to establish http connection to VPC-A from other two VPCs (kubernetes) it fails.

VPC-B and VPC-C runs microservices based application deployed on kubernetes(docker). So, it's not guaranteed that a micorservice pod will run exactly from specific instance. On re-deployment of the microservice, it jumps to any available instance in VPC.

Only when public IP of any instance from VPC-B or VPC-C added to security group of VPC-A instance , http request to VPC-A instance works from that specific instance of other VPCs. This can't be permanent solution due to possible instance expiry(and hence the IP) and nature of the application.

It was expected that setup will make it possible to access service running on instance in VPC-A from both of the other VPCs. Please point out what is missing or ill configured.