Update veda-deploy to accommodate non-core APIs for veda-routes

Context

veda-routes only configures CloudFront origins/behaviors for core VEDA APIs. Non-core APIs like titiler-cmr and titiler-multidim and features-api are manually configured, which means any redeployment of veda-routes on instances that include these services can misconfigure or remove their CloudFront behaviors.

Options

1. Add titiler-cmr and titiler-multidim as configurable origins in veda-routes (opt-in per instance)
2. Document which instances have non-core APIs and add deployment guardrails?
3. Move to a model where veda-routes reads its origin config from the instance's veda-deploy config

Acceptance Criteria

• Redeploying veda-routes does not remove or misconfigure non-core API origins
• Non-core API configuration is discoverable / documented?
• veda-deploy is updated to make veda-routes get features api url from veda-features-api step output

https://github.com/NASA-IMPACT/veda-deploy/blob/dev/.github/workflows/deploy.yml#L452-L461

      - name: Update Veda Routes secrets with dependency outputs
        shell: bash
        run: |
          python3 "${{ github.workspace }}/scripts/update_secret_with_inputs.py" \
            --secret-id ${{ vars.DEPLOYMENT_ENV_ROUTES_SECRET_NAME }} \
            --prefix=VEDA_ \
            --raster_api_url=${{ needs.deploy-veda-backend.outputs.raster_api_url }} \
            --ingest_api_url=${{ needs.deploy-veda-backend.outputs.ingest_api_url }} \
            --stac_api_url=${{ needs.deploy-veda-backend.outputs.stac_api_url }} \
            --stac_browser_bucket_name=${{ needs.deploy-veda-backend.outputs.stac_browser_bucket_name }}

Start with .github/workflows/deploy.yml lines 452-461 and scripts/update_secret_with_inputs.py to trace how deployment outputs become Veda Routes secrets. Review the listed options and the veda-features-api dependency output, then define the configuration and guardrails needed so redeployments preserve non-core API origins and expose the features API URL.