Even if the Kube sidecar starts very quickly, it's fully ready after ~2 minutes. It requires to find all the connections with other nodes.
If the Pod start to process the messages in the SQS queue, it can be that the first messages fail for a timeout.
Proposal:
- add a timaout of ~2 minutes when the pod start befor start to process the messages.
Even if the Kube sidecar starts very quickly, it's fully ready after ~2 minutes. It requires to find all the connections with other nodes.
If the Pod start to process the messages in the SQS queue, it can be that the first messages fail for a timeout.
Proposal: