Kubernetes Pod loses internet/local network connection on startup/restart, how to find error?

Question

SUSE

Hello everyone, there is a cluster: 3 masters, 4 workers
SUSE CaasP V3
Kubernetes v1.10.11

on it, in total, about 50 (including system) pods puff slightly, without loading
a couple of requests per second.
Last week, I connected/launched a replica for one of the pods (replicas: 2), it started up, but has no connection to the database (Postgres, also on the cluster) - I killed the replica without hesitation - restarted automatically - everything is fine and I did not give this value.

Today I added 6 new (new deployment) pods, two of them (different deployments) have the same problem.
Those. there is no connection to the base. Killed one (rebooted) - the same problem. Once again killed - it works.
I will support the second in order to find the problem!

I thought the problem was in the node - no (as I think). Other pods run on this worker, no problem.

The main problem is that there is no connection within the cluster, to the database .. But I can reach the very bottom from another (working) one. Those. It's all about outgoing communication.
For example, apt-get install nano also doesn't work:

Err:1 archive.ubuntu.com/ubuntu bionic/main amd64 nano amd64 2.9.3-2
Temporary failure resolving 'archive.ubuntu.com'
E: Failed to fetch archive.ubuntu.com/ubuntu/pool/main/n/ nano/nano_2.... Temporary failure resolving 'archive.ubuntu.com'
E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?

The problem arises in the pod itself, the container images are downloaded from the repository every time without any problems.

Any idea where to look for the problem?
Thank you!