[Telemetry docker] add memory and memory swap limits

[pra-moh]

Why I did it

Fix https://github.com/Azure/sonic-telemetry/issues/71

How I did it

Added memory limit for telemetry docker.
Historical docker memory usage shows telemetry docker consuming 150-200MB memory. Adding some extra buffer.

How to verify it

Build and install image with change:

docker memory for telemetry changes to assigned value:

Which release branch to backport (provide reason below if selected)

• 201811
• [- ] 201911
• [- ] 202006
• [- ] 202012

Description for the changelog

A picture of a cute animal (not mandatory but encouraged)

[rlhui]

@pra-moh - thanks, please add testing done.

[lguohan]

what is memory-swap?

[pra-moh]

what is memory-swap?

what is memory-swap?
Adding this argument to avoid container using any extra memory. Using just --memory argument can make docker start memory swap from disk based on some threads. This will make sure it will never use more than 300MB

Based on docker documentation:
PREVENT A CONTAINER FROM USING SWAP
If --memory and --memory-swap are set to the same value, this prevents containers from using any swap. This is because --memory-swap is the amount of combined memory and swap that can be used, while --memory is only the amount of physical memory that can be used.

https://docs.docker.com/config/containers/resource_constraints/#prevent-a-container-from-using-swap

[lguohan]

please describe what kind of testing have you done?

besides, we do not use swap at all on switches. it should be set to zero.

[lguohan]

can you update the description?

[lguohan]

can you update the description to say how 500m is derived?

[rlhui]

Let's revise the memory cap based on latest data. Thanks.

[lguohan]

bad pr, oom triggered system reboot.

[hui-ma]

let us revert it. heard about the issue after s/o.

[zhenggen-xu]

We need "--oom-kill-disable" option when we limit the memory so no kernel panic.

On our testing we found that as we move to Memory threshold CPU utilization start going over 100%+ by telemetry and also docker start getting very slow/unresponsive. High CPU will impact overall system.

Plan is to have monit itself check for telemetry docker memory threshold and if exceed this
We will alert and restart telemetry docker.