[Feature Request] Add optional Kubernetes Indexed Job template for external_batch simulation execution

Description

The openstudio-server project introduced the external_batch execution mode, which packages analysis datapoints into self-contained chunk archives (run_chunk.rb) to eliminate worker starvation and end-of-analysis compute valleys.

This issue proposes adding a new optional Kubernetes Job template (templates/job-external-batch.yaml) and corresponding configuration in values.yaml to allow running external batch chunks natively on Kubernetes alongside or in place of the standard Resque worker deployment, reusing the chart's existing NFS storage.

Motivation & Benefits

• Compute Valley Mitigation: Shifts job scheduling to Kubernetes, allowing analysis batches to overlap execution without bottlenecking the web-background Resque processes.
• Zero Extra Infrastructure: Reuses the existing Kubernetes cluster and Helm release without needing an external batch cluster (such as AWS Batch, SLURM, or Nomad).
• Dual-Mode Coexistence: Users can continue running traditional Resque worker deployments for iterative/standard analyses (batch_run, ga, pso) while executing external_batch_run workloads on-demand via indexed Kubernetes Jobs.
• Storage Reuse: Job pods mount the existing shared NFS Persistent Volume Claim (ReadWriteMany), avoiding duplicate data transfer.

Proposed Changes

1. values.yaml additions

Add a dedicated configuration section for external_batch:

external_batch:
  enabled: false
  analysis_id: ""
  completions: 1       # Total number of chunks generated by the packager
  parallelism: 1       # Concurrency / number of chunk pods running simultaneously
  backoff_limit: 3
  active_deadline_seconds: 86400
  ttl_seconds_after_finished: 3600

  container:
    image: nrel/openstudio-server:latest
    imagePullPolicy: IfNotPresent

  storage:
    mountPath: /shared-data
    # Reuses existing chart NFS PVC if left blank
    existingClaim: ""

  resources:
    requests:
      cpu: "2000m"
      memory: "4Gi"
    limits:
      cpu: "4000m"
      memory: "8Gi"

**2. New Template: templates/job-external-batch.yaml**

Leverage Kubernetes completionMode: Indexed so the environment variable JOB_COMPLETION_INDEX dynamically matches the chunk ID:

{{- if .Values.external_batch.enabled }}
apiVersion: batch/v1
kind: Job
metadata:
  name: {{ include "openstudio-server.fullname" . }}-batch-{{ .Values.external_batch.analysis_id | default "manual" }}
  labels:
    app.kubernetes.io/name: {{ include "openstudio-server.name" . }}-external-batch
    app.kubernetes.io/instance: {{ .Release.Name }}
    app.kubernetes.io/component: simulation-job
spec:
  completions: {{ .Values.external_batch.completions }}
  parallelism: {{ .Values.external_batch.parallelism }}
  completionMode: Indexed
  backoffLimit: {{ .Values.external_batch.backoff_limit }}
  activeDeadlineSeconds: {{ .Values.external_batch.active_deadline_seconds }}
  ttlSecondsAfterFinished: {{ .Values.external_batch.ttl_seconds_after_finished }}
  template:
    metadata:
      labels:
        app.kubernetes.io/name: {{ include "openstudio-server.name" . }}-external-batch
        app.kubernetes.io/instance: {{ .Release.Name }}
    spec:
      restartPolicy: OnFailure
      containers:
        - name: chunk-runner
          image: {{ .Values.external_batch.container.image | default .Values.worker.container.image }}
          imagePullPolicy: {{ .Values.external_batch.container.imagePullPolicy }}
          command: ["/bin/bash", "-c"]
          args:
            - |
              ruby /openstudio-server/external_batch/runner/run_chunk.rb \
                --analysis-id {{ .Values.external_batch.analysis_id }} \
                --chunk-id ${JOB_COMPLETION_INDEX} \
                --storage-dir {{ .Values.external_batch.storage.mountPath }}
          resources:
            {{- toYaml .Values.external_batch.resources | nindent 12 }}
          volumeMounts:
            - name: shared-nfs
              mountPath: {{ .Values.external_batch.storage.mountPath }}
      volumes:
        - name: shared-nfs
          persistentVolumeClaim:
            claimName: {{ .Values.external_batch.storage.existingClaim | default (printf "%s-nfs" (include "openstudio-server.fullname" .)) }}
{{- end }}

Usage Example (Ad-Hoc Execution)

Users can template and dispatch batch jobs on demand for a packaged analysis without modifying the running release:

helm template openstudio-server ./openstudio-server-helm \
  -s templates/job-external-batch.yaml \
  --set external_batch.enabled=true \
  --set external_batch.analysis_id="<ANALYSIS_UUID>" \
  --set external_batch.completions=16 \
  --set external_batch.parallelism=8 | kubectl apply -f -

Acceptance Criteria

• external_batch values block defined in values.yaml (default enabled: false).
• templates/job-external-batch.yaml added using completionMode: Indexed.
• Job pods successfully mount the existing NFS PVC and access analysis chunk packages.
• Unit tests added verifying helm template renders the Job only when enabled.
• Documentation added to README.md explaining dual-mode operation and external batch dispatch.