Support LNC with trn2

[movence]

LNC (Logical Neuron Cores), concept to represents multiple Neuron cores as a single Neuron core, is a new feature supported with trn2. This requires a new volume mount for nueron-monitor to get LNC configuration.

Description of changes:

• Bump neuron-monitor image to 1.3.0
• Add new volume mount for /opt
• Update GOMEMLIMIT to 320MiB

Tested by deploying the changes to a test cluster:

# describe neuron monitor (some truncated)

Containers:
  neuron-monitor:
    Image:         public.ecr.aws/neuron/neuron-monitor:1.3.0
    Image ID:      public.ecr.aws/neuron/neuron-monitor@sha256:[sha]
    Port:          8000/TCP
    Host Port:     0/TCP
    ...
    State:          Running
      Started:      Thu, 09 Jan 2025 13:03:44 -0500
    Ready:          True
    Limits:
      cpu:     500m
      memory:  256Mi
    Requests:
      cpu:     256m
      memory:  128Mi
    Environment:
      NODE_NAME:    (v1:spec.nodeName)
      PATH:        /usr/local/bin:/usr/bin:/bin:/opt/aws/neuron/bin
      GOMEMLIMIT:  320MiB
    Mounts:
      /etc/amazon-cloudwatch-observability-neuron-cert/ from neurontls (ro)
      /etc/neuron-monitor-config from neuron-monitor-config (rw)
      /opt-aws from aws-config (ro)
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-hxmsx (ro)
    Volumes:
      aws-config:
        Type:          HostPath (bare host directory volume)
        Path:          /opt/aws
        HostPathType:

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[mounchin]

Can you also review this CR
https://github.com/aws-observability/helm-charts/pull/136/files

[sky333999]

The default limit itself is lower than this -

helm-charts/charts/amazon-cloudwatch-observability/values.yaml

Line 1432 in 2711d52

memory: 256Mi

Should we updating that as well?