--- name: coreweave-multi-env-setup-64f6bf15b67c description: 'Configure CoreWeave across development, staging, and production environments. Use when setting up multi-environment GPU infrastructure, separating namespaces, or managing per-environment GPU quotas. Trigger with phrases like "coreweave environments", "coreweave staging", "coreweave multi-env", "coreweave namespace setup". ' allowed-tools: Read, Write, Edit, Bash(kubectl:*), Bash(kustomize:*), Grep version: 1.11.0 license: MIT author: Jeremy Longshore tags: - saas - gpu-cloud - kubernetes - inference - coreweave compatibility: Designed for Claude Code --- # CoreWeave Multi-Environment Setup > **Community-contributed.** Not affiliated with, endorsed by, or sponsored by CoreWeave, Inc. CoreWeave is a registered trademark of CoreWeave, Inc. ## Overview CoreWeave GPU cloud requires strict environment separation to control infrastructure costs and prevent resource contention. Each environment maps to an isolated Kubernetes namespace with its own GPU quota, scaling policy, and access controls. Development uses cheaper GPU tiers for iteration speed, staging mirrors production GPU types for accurate benchmarking, and production runs full-scale with no scale-to-zero to guarantee inference latency SLAs. ## Environment Configuration ## Prerequisites - Separate, approved namespaces and service identities for development, staging, and production. - Environment-specific secrets injected by a secrets manager, never committed `.env` files. - A promotion owner, staging evaluation gate, and production rollback manifest. ## Instructions 1. Define overlays that differ only in reviewed capacity, endpoint, and namespace values. 2. Validate required variables and secret references before applying an overlay. 3. Promote dev to staging, run the service evaluation, then use a controlled production rollout. 4. Roll back the production overlay on an SLO, quality, or security-gate failure; do not copy staging credentials into production. ```typescript const coreweaveConfig = (env: string) => ({ development: { namespace: "app-dev", apiEndpoint: process.env.CW_API_ENDPOINT_DEV!, token: process.env.CW_TOKEN_DEV!, gpuType: "L40", scaleToZero: true, replicas: [0, 1], }, staging: { namespace: "app-staging", apiEndpoint: process.env.CW_API_ENDPOINT_STG!, token: process.env.CW_TOKEN_STG!, gpuType: "A100_PCIE_40GB", scaleToZero: true, replicas: [0, 2], }, production: { namespace: "app-prod", apiEndpoint: process.env.CW_API_ENDPOINT_PROD!, token: process.env.CW_TOKEN_PROD!, gpuType: "A100_PCIE_80GB", scaleToZero: false, replicas: [2, 10], }, }[env]); ``` ## Environment Files ```text # Per-env files: .env.development, .env.staging, .env.production CW_API_ENDPOINT_{DEV|STG|PROD}=https://k8s.{ord1|ord1|las1}.coreweave.com CW_TOKEN_{DEV|STG|PROD}= CW_NAMESPACE={app-dev|app-staging|app-prod} CW_GPU_TYPE={L40|A100_PCIE_40GB|A100_PCIE_80GB} ``` ## Environment Validation ```typescript function validateCoreWeaveEnv(env: string): void { const required = ["CW_API_ENDPOINT", "CW_TOKEN", "CW_NAMESPACE", "CW_GPU_TYPE"]; const suffix = { development: "_DEV", staging: "_STG", production: "_PROD" }[env]; const missing = required .map((k) => (k.includes("NAMESPACE") ? k : `${k}${suffix}`)) .filter((k) => !process.env[k]); if (missing.length) throw new Error(`Missing env vars for ${env}: ${missing.join(", ")}`); } ``` ## Promotion Workflow ```bash # 1. Validate model in dev namespace kubectl -n app-dev get inferenceservice my-model -o jsonpath='{.status.conditions}' # 2. Apply staging overlay with production GPU type kustomize build k8s/overlays/staging | kubectl apply -f - # 3. Run inference benchmarks against staging endpoint curl -X POST https://staging.myapp.coreweave.cloud/v1/predict -d @test-payload.json # 4. Promote to production (blue-green via namespace switch) kustomize build k8s/overlays/prod | kubectl apply -f - kubectl -n app-prod rollout status deployment/my-model ``` ## Environment Matrix | Setting | Dev | Staging | Prod | |---------|-----|---------|------| | GPU Type | L40 | A100 40GB | A100 80GB | | Scale-to-Zero | Yes | Yes | No | | Replicas | 0-1 | 0-2 | 2-10 | | Namespace | app-dev | app-staging | app-prod | | Region | ord1 | ord1 | las1 | | Spot Instances | Yes | No | No | ## Error Handling | Issue | Cause | Fix | |-------|-------|-----| | GPU quota exceeded | Namespace limit reached | Request quota increase via CW support portal | | Pod stuck Pending | GPU type unavailable in region | Check `kubectl describe node` for capacity; switch region | | Scale-to-zero not waking | HPA misconfigured | Verify `minReplicas: 0` and KEDA scaler settings | | Namespace access denied | RBAC not applied to overlay | Apply `RoleBinding` in kustomize overlay | ## Output - Environment-isolated manifests, identities, and quotas with a documented promotion path. - A redacted validation and rollout receipt for each environment. - A production rollback path that preserves the prior known-good revision. ## Examples Validate a staging overlay server-side before rollout, then wait for its deployment: ```bash kustomize build k8s/overlays/staging | kubectl apply --dry-run=server -f - kustomize build k8s/overlays/staging | kubectl apply -f - kubectl -n app-staging rollout status deployment/my-model --timeout=10m ``` If the staging result fails its signed gate, stop promotion and restore the prior staging revision. Keep credentials out of terminal history, logs, and overlay files. ## Resources - [CoreWeave CKS](https://docs.coreweave.com/docs/products/cks) ## Next Steps See `coreweave-deploy-integration`. --- Source attribution: jeremylongshore Original project: https://github.com/jeremylongshore/claude-code-plugins-plus-skills/blob/ab85a061f7bcc039dac13ee193de7a4cd4388571/plugins/saas-packs/coreweave-pack/skills/coreweave-multi-env-setup/SKILL.md License: MIT License information: https://opensource.org/license/mit Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice MIT License Copyright (c) 2025 Jeremy Longshore Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. # Mega AI Hub resource metadata Creator / publisher: jeremylongshore Original project: https://github.com/jeremylongshore/claude-code-plugins-plus-skills/blob/ab85a061f7bcc039dac13ee193de7a4cd4388571/plugins/saas-packs/coreweave-pack/skills/coreweave-multi-env-setup/SKILL.md Attribution basis: Source repository record Fields of work: Development & Engineering Roles: Backend Developer, DevOps / SRE, ML / AI Engineer Specialties: API, Design, DevOps, Frontend, Testing Tools: Kubernetes, Coreweave Country applicability: Country not specified Classification basis: Source catalog labels Instruction SHA-256: 60b43aefbeebf78ef4b3739000a9c5b8f93c610676055937829a5ab7d26de480