eks-autoscaling-automator
Designs and implements EKS cluster autoscaling automation. Compares and deploys Cluster Autoscaler vs Karpenter (provisioner node templates, consolidation policies, disruption budgets), configures node-group scaling policies, builds HPA (Horizontal Pod Autoscaler) with custom and external metrics, explains VPA (Vertical Pod Autoscaler) caveats and incompatibilities, deploys KEDA for event-driven scaling (Kafka, SQS, Prometheus), handles spot instance interruption (termination handler, graceful drain), configures overprovisioning pause-pods for fast scale-up, sets up pod disruption budgets to protect quorum, defines priority classes for workload preemption, and deploys descheduler for bin-packing optimization. Emits AUTOMATION_DEPLOYED with ready-to-apply Helm values + manifests or REVIEW_REQUIRED with the specific gap. Use when building EKS autoscaling, choosing Karpenter vs Cluster Autoscaler, or optimizing an existing scaling setup.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-24
- Stars
- ★ 4
- Forks
- 0
01 前言 Front Matter
- name
- eks-autoscaling-automator
- license
- Apache-2.0
- metadata
- { "tags": "eks, karpenter, cluster-autoscaler, hpa, keda, spot, autoscaling, automate", "phase": "4", "author": "Jacky Chan — AWS Community Builder", "domain": "aws-cloudops", "family": "Compute", "version": "0.1.0", "keywords": "EKS autoscaling, Karpenter, Cluster Autoscaler, Horizontal Pod Autoscaler, HPA, Vertical Pod Autoscaler, VPA, KEDA, spot instance interruption, overprovisioning, pod disruption budget, priority class, descheduler, consolidation, disruption budget", "task_type": "automate", "complexity": "high", "entry_point": "false", "skill_class": "capability", "when_to_use": "Building EKS autoscaling (Karpenter or Cluster Autoscaler), configuring HPA/VPA/KEDA for workload scaling, handling spot instance interruptions, optimizing scale-up latency with overprovisioning, protecting workloads with PDBs and priority classes, or improving bin-packing with descheduler.", "requires_llm": "true", "verdict_shape": "AUTOMATION_DEPLOYED | REVIEW_REQUIRED", "lifecycle_status": "active", "invocation_schema": "Input: either (a) an EKS cluster description (version, node count, workload mix, scaling requirements, spot/on-demand ratio) plus observability metrics, OR (b) an autoscaling request (\"set up Karpenter with spot diversification\", \"configure HPA with custom metrics for the API deployment\"). Output: deterministic AUTOSCALING block — TOOLING/PROVISIONING/POD_SCALING/SPOT/SAFETY/VERDICT — where VERDICT is AUTOMATION_DEPLOYED (Helm values + manifests ready) or REVIEW_REQUIRED (specific gap cited).", "supports_pipeline": "true", "activation_triggers": "EKS autoscaling, Karpenter setup, Cluster Autoscaler, horizontal pod autoscaler, KEDA event-driven scaling, spot interruption handling, overprovisioning pause pods, pod disruption budget, descheduler bin-packing, node group scaling policy" }
- description
- Designs and implements EKS cluster autoscaling automation. Compares and deploys Cluster Autoscaler vs Karpenter (provisioner node templates, consolidation policies, disruption budgets), configures node-group scaling policies, builds HPA (Horizontal Pod Autoscaler) with custom and external metrics, explains VPA (Vertical Pod Autoscaler) caveats and incompatibilities, deploys KEDA for event-driven scaling (Kafka, SQS, Prometheus), handles spot instance interruption (termination handler, graceful drain), configures overprovisioning pause-pods for fast scale-up, sets up pod disruption budgets to protect quorum, defines priority classes for workload preemption, and deploys descheduler for bin-packing optimization. Emits AUTOMATION_DEPLOYED with ready-to-apply Helm values + manifests or REVIEW_REQUIRED with the specific gap. Use when building EKS autoscaling, choosing Karpenter vs Cluster Autoscaler, or optimizing an existing scaling setup.
- compatibility
- Agent runtime that reads SKILL.md (Claude Code, Cursor, Windsurf, Codex, Gemini). No AWS CLI or kubectl required for offline workflow design. Live deployment uses helm install/upgrade (Cluster Autoscaler, Karpenter, KEDA, descheduler), kubectl apply (HPA, VPA, PDB, priority classes, Karpenter NodePool/NodeClaim), aws eks update-nodegroup-config (managed node group scaling) — AWS CLI v2, kubectl, helm v3.
02 资源文件 20
- 参考文档eval/test-cases.yaml6.5 KB
- 参考文档evals/baselines/ca-to-karpenter-migration.md892 B
- 参考文档evals/baselines/karpenter-spot-full-deploy.md1.3 KB
- 参考文档evals/baselines/keda-sqs-scale-to-zero.md1.0 KB
- 参考文档evals/baselines/missing-metrics-server.md750 B
- 参考文档evals/baselines/vpa-hpa-conflict.md929 B
- 参考文档evals/eval_config.yaml1.7 KB
- 参考文档evals/evals.json4.1 KB
- 参考文档evals/prompts/ca-to-karpenter-migration.md760 B
- 参考文档evals/prompts/karpenter-spot-full-deploy.md959 B
- 参考文档evals/prompts/keda-sqs-scale-to-zero.md672 B
- 参考文档evals/prompts/missing-metrics-server.md483 B
- 参考文档evals/prompts/vpa-hpa-conflict.md518 B
- 参考文档examples/end-to-end.md9.5 KB
- 参考文档references/advanced-patterns.md5.3 KB
- 参考文档references/diagnostic-commands.md643 B
- 参考文档references/hpa-keda-descheduler.md11.0 KB
- 参考文档references/karpenter-configuration.md8.1 KB
- 参考文档references/worked-examples.md648 B
- 参考文档skill.usf.yaml4.5 KB
03 SKILL.md 正文
EKS Autoscaling Automator
Mindset
One-line takeaway: every EKS autoscaling system operates at two layers — cluster autoscaling (adding/removing nodes: Karpenter or Cluster Autoscaler) and pod autoscaling (adding/removing pods: HPA or KEDA). A gap in either layer produces a silent failure: pods are Pending because no node has capacity, or nodes sit idle because pod count never scales with load.
- Cluster Autoscaler and Karpenter are mutually exclusive. Running both causes conflicting scaling decisions — CA provisions a node, Karpenter de-provisions it, ad infinitum.
- HPA without custom metrics is just CPU autoscaling. For request- driven workloads (latency, queue depth), HPA needs custom/external metrics via Prometheus Adapter or KEDA.
- Spot without interruption handling is an outage waiting to happen. Spot instances get a 2-minute warning. Without the termination handler or Karpenter's native disruption, pods are killed ungracefully.
Quick navigation
| You want to... | Go to |
|---|---|
| Choose Karpenter vs Cluster Autoscaler | Step 1 |
| Deploy Karpenter with NodePool + disruption budget | Step 2 |
| Deploy Cluster Autoscaler with auto-discovery | Step 3 |
| Configure managed node group scaling | Step 4 |
| Set up HPA with custom metrics | Step 5 |
| Understand VPA caveats | Step 6 |
| Deploy KEDA for event-driven scaling | Step 7 |
| Handle spot interruptions | Step 8 |
| Configure overprovisioning pause-pods | Step 9 |
| Set up PDBs and priority classes | Step 10 |
| Deploy descheduler for bin-packing | Step 11 |
Critical rules at a glance (do NOT bury these)
-
Karpenter consolidation replaces nodes aggressively.
WhenEmptyOrUnderutilizedwill DELETE underutilized nodes and reschedule pods. For production, usedisruptionBudgetsto cap how many nodes Karpenter disrupts at once. Without a budget, Karpenter can drain 50% of nodes simultaneously. -
Spot diversification across instance families is mandatory. A NodePool requesting only
c5.largespot will be evicted en masse when c5 capacity disappears. Diversify across at least 3-4 families (c5, m5, c6i, m6i). -
Overprovisioning pause-pods create headroom for fast scale-up. Without them, new pods wait 2-5 minutes for a node to boot. Pause- pods (low priority) occupy scheduled capacity; real workloads preempt them and schedule immediately.
-
HPA and VPA are incompatible for the same CPU/Memory dimension. HPA scales replicas based on CPU; VPA adjusts CPU requests. Running both on CPU causes oscillation. Use VPA in
Offmode alongside HPA, or VPA for memory only. -
PDBs protect against voluntary disruption only. A PDB with
minAvailable: 2preventskubectl drainfrom evicting below 2 replicas. It does NOT protect against spot termination (involuntary). For spot, combine PDB + termination handler + topology spread.
Pre-flight: data requirements
| Input | Source | Why |
|---|---|---|
| EKS cluster version | eks describe-cluster |
Karpenter v1 requires 1.27+ |
| Current node count/type | kubectl get nodes -o wide |
Baseline fleet |
| Workload resource requests | kubectl get pods -A -o json |
Drives node sizing |
| Current HPA/VPA/KEDA | kubectl get hpa,vpa -A |
Don't duplicate |
| Spot vs on-demand ratio | Business requirements | Cost target |
| Metrics server health | kubectl top pods |
Required for default HPA |
| Prometheus (if custom metrics) | kubectl get svc -n monitoring |
For custom-metric HPA |
If the input is malformed, emit:
AUTOSCALING: <reference>
VERDICT: ERROR
REASON: Cannot design autoscaling — cluster version, node count, and workload requests are required.
GAP: Run kubectl get nodes -o wide and kubectl get pods -A -o json.
Process — Autoscaling design (apply in order)
Step 0: Expert knowledge — non-obvious autoscaling behaviors
Step 0 expert-knowledge deep dive (Karpenter v1 API migration, consolidation modes, CA scale-down blockers, KEDA trigger semantics, IMDS spot notices, pause-pod priority, VPA Auto restarts, descheduler PodLifeTime, MNG desiredSize override) moved verbatim to references/advanced-patterns.md. Load on demand before choosing or tuning any autoscaling component.
Step 1: Choose Karpenter vs Cluster Autoscaler
| Dimension | Karpenter | Cluster Autoscaler |
|---|---|---|
| Scaling speed | ~30s (direct EC2 API) | ~2-5 min (ASG lifecycle) |
| Node selection | Flexible (any type, AZ, spot+OD mix) | Rigid (ASG instance types) |
| Consolidation | Yes (deletes underutilized nodes) | Limited (empty nodes only) |
| Spot handling | Native disruption + diversified | Via ASG mixed policy |
| Maturity | v1 GA (2024) | Battle-tested, stable |
| Best for | Dynamic, spot-heavy, fast scale | Stable, predictable capacity |
Choose Karpenter if: spot priority, < 1 min scale-up, consolidation needed, heterogeneous types. Choose CA if: fixed types in managed node groups, simplicity priority.
NEVER run both. Migrate: uninstall CA first, then install Karpenter.
Step 2: Deploy Karpenter with NodePool and disruption budgets
helm install karpenter oci://public.ecr.aws/karpenter/karpenter \
--namespace kube-system \
--set clusterName=my-prod-cluster \
--set interruptionQueue=true \
--version 1.0.0
NodePool with spot diversification + disruption budget:
apiVersion: karpenter.sh/v1
kind: NodePool
metadata:
name: default
spec:
template:
spec:
requirements:
- key: karpenter.k8s.aws/instance-category
operator: In
values: ["c", "m"]
- key: karpenter.k8s.aws/instance-generation
operator: Gt
values: ["5"]
- key: karpenter.sh/capacity-type
operator: In
values: ["spot", "on-demand"]
nodeClassRef:
group: karpenter.k8s.aws
kind: EC2NodeClass
name: default
limits:
cpu: 1000
disruption:
consolidationPolicy: WhenEmptyOrUnderutilized
consolidateAfter: 30s
budgets:
- nodes: "20%"
- schedule: "0 9 * * Mon-Fri"
duration: 8h
nodes: "0"
Full EC2NodeClass manifest (AMI alias, subnet/SG selectors, role, block devices) moved verbatim to references/karpenter-configuration.md. Load on demand when authoring the EC2NodeClass that pairs with the NodePool above.
Key configuration:
disruption.budgets: First budget allows 20% disruption normally. Second budget blocks ALL disruption Mon-Fri 09:00-17:00. Critical for production stability.- Spot diversification:
instance-category: [c, m]+generation: >5allows 8+ instance types. If c5 spot disappears, Karpenter tries m5 or c6i immediately. interruptionQueue: true: enables native spot interruption handling via SQS + EventBridge. Do NOT install NTH when this is enabled.
See references/karpenter-configuration.md for full API reference, requirement keys, and consolidation troubleshooting.
Step 3: Deploy Cluster Autoscaler with auto-discovery
helm install cluster-autoscaler autoscaler/cluster-autoscaler \
--namespace kube-system \
--set autoDiscovery.clusterName=my-prod-cluster \
--set awsRegion=us-east-1 \
--version 9.34.0
Auto-discovery tags (must be on node group ASGs):
k8s.io/cluster-autoscaler/my-prod-cluster = owned
k8s.io/cluster-autoscaler/enabled = true
Key tuning: scaleDownUnneededTime: "30m",
scaleDownUtilizationThreshold: "0.5",
skipNodesWithLocalStorage: true, expander: "least-waste".
Step 4: Configure managed node group scaling
aws eks update-nodegroup-config \
--cluster-name my-prod-cluster --nodegroup-name my-ng \
--scaling-config desiredSize=5,minSize=3,maxSize=20
For EKS Auto Mode (2025+, simplest path for new clusters):
aws eks update-cluster-config --name my-prod-cluster \
--compute-config '{"nodeRoleArn":"arn:...","enabled":true}'
Step 5: Set up HPA with custom metrics
HPA with CPU + custom metric + behavior tuning:
Full HPA manifest (CPU + custom metric + scaleUp/scaleDown behavior policies) moved verbatim to references/hpa-keda-descheduler.md. Load on demand when authoring the HPA object for a deployment.
Behavior tuning guide: scale UP fast (users waiting), scale DOWN
slow (avoid thrashing). scaleDown.stabilizationWindowSeconds: 300
minimum for production.
See references/hpa-keda-descheduler.md for custom metrics setup, KEDA triggers, and descheduler strategies.
Step 6: VPA caveats and incompatibilities
| Mode | Behavior | Production-safe? |
|---|---|---|
Auto |
Recreates pods to apply requests | No (causes restarts) |
Initial |
Sets requests only on NEW pods | Yes |
Off |
Recommendations only | Yes |
HPA + VPA CPU conflict: VPA raises CPU request → HPA sees lower
utilization → HPA scales down → loop. Resolution: VPA Off mode (apply
manually) or VPA for memory only.
Step 7: Deploy KEDA for event-driven scaling
helm install keda kedacore/keda --namespace keda-system --create-namespace --version 2.15.0
SQS-triggered ScaledObject (scale-to-zero):
SQS-triggered ScaledObject manifest (scale-to-zero, queueLength trigger, polling/cooldown) moved verbatim to references/hpa-keda-descheduler.md. Load on demand when authoring a KEDA ScaledObject.
KEDA vs HPA: use KEDA when you need scale-to-zero, external-system triggers, or multiple trigger sources. Use HPA when CPU/Memory suffices.
Step 8: Handle spot instance interruptions
Karpenter approach (recommended): Set interruptionQueue: true.
Karpenter watches SQS fed by EventBridge, receives 2-min notice, cordons
- drains + provisions replacement. No additional components.
Cluster Autoscaler approach (NTH):
helm install aws-node-termination-handler eks/aws-node-termination-handler \
--namespace kube-system \
--set enableSpotInterruptionDraining=true \
--set enableQueueProcessing=true \
--set awsRegion=us-east-1
NEVER run both — they conflict on the same SQS queue.
Spot diversification rules:
| Rule | Why |
|---|---|
| 3+ instance families | Reduces mass-reclamation risk |
| 2+ instance sizes | Same |
| 2+ AZs | Capacity shortages are AZ-specific |
capacity-type: [spot, on-demand] |
Fallback to on-demand |
Step 9: Configure overprovisioning pause-pods
Priority class (MUST be lower than default):
apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
name: overprovisioning
value: -1
description: "Pause-pod priority (lowest)."
Pause-pod Deployment:
apiVersion: apps/v1
kind: Deployment
metadata:
name: pause-pods
namespace: kube-system
spec:
replicas: 10
strategy: {type: Recreate}
selector: {matchLabels: {app: pause-pod}}
template:
metadata: {labels: {app: pause-pod}}
spec:
priorityClassName: overprovisioning
containers:
- name: pause
image: registry.k8s.io/pause:3.9
resources:
requests: {cpu: 1, memory: 1Gi}
Each pause-pod occupies 1 CPU + 1 GiB. For 10 pods headroom, Karpenter provisions 10 pods worth of extra nodes. When a real deployment scales up, it preempts pause-pods and schedules immediately.
Step 10: Pod disruption budgets and priority classes
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
name: api-server-pdb
spec:
minAvailable: 2
selector: {matchLabels: {app: api-server}}
Priority classes:
# production-critical: value 1000000
# production-standard: value 100000
# batch-default: value 10000
# default (globalDefault): value 0
# overprovisioning: value -1
Preemption flow: when a node is full, the scheduler checks if lower- priority pods can be evicted to make room for a higher-priority pod.
Step 11: Deploy descheduler for bin-packing
helm install descheduler descheduler/descheduler \
--namespace kube-system \
--set kind=Deployment \
--set schedule="0 */2 * * *" \
--set deschedulingInterval=2h \
--version 0.31.0
Key safety settings: maxNoOfPodsToEvictPerNode: 3,
maxNoOfPodsToEvictTotal: 50, evictSystemCriticalPods: false,
nodeFit: true (CRITICAL — prevents evicting pods with nowhere to go).
Run in DryRun mode first: --set dryDescheduling=true.
Output format
AUTOSCALING: <reference>
CLUSTER: <EKS version + node count>
TOOLING:
- Cluster autoscaler: Karpenter <ver> | CA <ver> | EKS Auto Mode
- Pod autoscaler: HPA | KEDA | VPA (Off)
- Metrics: metrics-server | Prometheus Adapter
PROVISIONING:
- NodePool: <name> / Instance types: <diversified list>
- Spot ratio: <%> / Disruption budget: <nodes per window>
POD_SCALING:
- HPA: <deployment> min=N max=N metric=<type>
- KEDA: <deployment> trigger=<type> min=N max=N
- VPA: <deployment> mode=<Off|Initial>
SPOT:
- Interruption handling: Karpenter native | NTH
- Diversification: <families> / Graceful drain: enabled|disabled
SAFETY:
- Overprovisioning: <N> x <cpu/mem>
- PDB: <deployment> minAvailable=N
- Priority classes: <list> / Descheduler: enabled|disabled
VERDICT: AUTOMATION_DEPLOYED | REVIEW_REQUIRED
GAP: <if REVIEW_REQUIRED, the specific missing piece>
TEMPLATE: <Helm values + kubectl manifests>
Worked example — AUTOMATION_DEPLOYED
AUTOSCALING: prod-cluster-autoscaling
CLUSTER: EKS 1.30, 25 nodes
TOOLING: Karpenter 1.0.0 (interruptionQueue) | HPA + KEDA | metrics-server + Prom Adapter
PROVISIONING: default NodePool / c5,m5,c6i,m6i / 70% spot 30% OD / 20% budget, freeze biz hrs
POD_SCALING: HPA api-server 3-50 (CPU+custom) | KEDA sqs-consumer 1-30 (SQS) | VPA Off
SPOT: Karpenter native / 4 families / graceful drain
SAFETY: 10 pause-pods x 1CPU/1Gi / PDB api-server minAvail=2 / 4 priority classes / descheduler on
VERDICT: AUTOMATION_DEPLOYED
GAP: None
Worked example — REVIEW_REQUIRED
REVIEW_REQUIRED worked example (staging cluster with metrics-server missing, no HPA/PDBs, on-demand only) moved verbatim to references/worked-examples.md. Load on demand when the verdict is REVIEW_REQUIRED; the primary AUTOMATION_DEPLOYED example stays inline.
Anti-Patterns — NEVER do these things
-
NEVER run Karpenter and Cluster Autoscaler simultaneously. They conflict: CA provisions, Karpenter deletes, CA re-provisions.
-
NEVER use
WhenEmptyOrUnderutilizedwithout a disruption budget in production. Karpenter can simultaneously disrupt a large fraction of nodes during consolidation. -
NEVER use spot from a single instance family. Diversify across at least 3-4 families to survive capacity reclamation events.
-
NEVER run NTH alongside Karpenter's interruption handling. Both watch the same SQS queue, causing double-drain race conditions.
-
NEVER set HPA
scaleDown.stabilizationWindowSeconds: 0. Zero causes immediate scale-down on transient dips, then scale-back-up — thrashing. Set at least 300s. -
NEVER use VPA
Autoon single-replica Deployments. VPA recreates the pod, causing an outage. UseOfforInitial. -
NEVER run HPA on CPU AND VPA on CPU simultaneously. They oscillate: VPA raises request → HPA sees lower utilization → HPA scales down.
-
NEVER set pause-pod priority equal to or higher than real workloads. Must be lower (e.g.,
-1) so real pods can preempt them. -
NEVER omit PDBs for multi-replica production workloads. Without PDB, voluntary disruption can evict all replicas simultaneously.
-
NEVER run descheduler without
maxNoOfPodsToEvictPerNodeandmaxNoOfPodsToEvictTotal. Without caps, mass eviction overwhelms the scheduler. Start with 3/node and 50 total. -
NEVER set CA
scaleDownUtilizationThresholdabove 0.7. Too high prevents scale-down; too low causes churn. 0.5 is the sweet spot. -
NEVER rely on PDB alone for spot workloads. PDB protects against voluntary disruption only. Combine with Karpenter/NTH + topology spread constraints.
-
NEVER enable KEDA without
cooldownPeriod. Default is 300s; setting to 0 causes rapid scale-up/scale-down oscillation for bursty workloads.
Pre-flight safety checks
Pre-flight safety checks (CONFIRM gate, CA-to-Karpenter migration, WhenEmptyOrUnderutilized graduation, descheduler DryRun, spot SQS verification) moved verbatim to references/diagnostic-commands.md. Load on demand before executing any state-changing autoscaling operation.
Appendix A — Karpenter vs CA decision tree
Karpenter-vs-CA decision tree appendix moved verbatim to references/advanced-patterns.md; the Step 1 comparison table stays inline. Load on demand when deciding the cluster-autoscaler tool.
Appendix B — HPA behavior quick reference
HPA behavior quick-reference appendix (API/worker/batch scale policies) moved verbatim to references/advanced-patterns.md. Load on demand when tuning HPA behavior policies.
Appendix C — Spot instance diversification
Spot diversification appendix (per-workload family tables + Karpenter flexible selector) moved verbatim to references/advanced-patterns.md. Load on demand when planning spot instance diversification.
Recent AWS features (2024-2026)
Recent AWS features (Karpenter v1 GA, disruption budgets, EKS Auto Mode, KEDA v2.15, HPA v2 behavior, VPA in-place resize) moved verbatim to references/advanced-patterns.md. Load on demand when picking component versions.
Expert heuristic: autoscaling blast radius
Autoscaling blast-radius heuristic (conservative-first rule, component interaction cascade, 4-phase wave-enabling pattern, failure-detection alarms) moved verbatim to references/advanced-patterns.md. Load on demand when planning phased rollout or setting failure alarms.
References (load on demand)
- references/advanced-patterns.md — Step 0 expert behaviours, Appendices A-C, Recent AWS features (2024-2026), and the blast-radius heuristic, moved from SKILL.md
- references/worked-examples.md — secondary REVIEW_REQUIRED worked example, moved from SKILL.md (primary AUTOMATION_DEPLOYED example stays inline)
- references/diagnostic-commands.md — pre-flight safety checks, moved from SKILL.md
- references/karpenter-configuration.md — Karpenter API reference — extended with the EC2NodeClass manifest moved from SKILL.md
- references/hpa-keda-descheduler.md — HPA/KEDA/descheduler detail — extended with the HPA and KEDA ScaledObject manifests moved from SKILL.md
Domain
AWS CloudOps / Compute Automation — EKS autoscaling.
AWS documentation
- Karpenter — https://karpenter.sh/docs/
- Cluster Autoscaler on EKS — https://docs.aws.amazon.com/eks/latest/userguide/cluster-autoscaler.html
- Horizontal Pod Autoscaler — https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/
- KEDA — https://keda.sh/docs/
- EKS Auto Mode — https://docs.aws.amazon.com/eks/latest/userguide/auto-mode.html
04 同仓库更多 Skills
xray-tracing-deployer
Deploys AWS X-Ray distributed tracing with production-grade configuration: X-Ray daemon deployment (EC2 systemd, ECS sidecar, EKS DaemonSet, Lambda built-in), SDK instrumentation (Java, Python, Node.js, Go, .NET) with annotations and metadata, sampling rules (default reservoir + rate vs custom rules with URL / method / service-name predicates), service map generation, trace query and retrieval, groups and insights, CloudWatch ServiceLens integration, IAM permissions (PutTraceSegments, GetSamplingRules), and latest features (Lambda Powertools tracing, OpenTelemetry / ADOT Collector support, W3C trace context). Emits a READY_TO_DEPLOY checklist with every configuration item verified. Use when enabling X-Ray on a new application, adding tracing to ECS / EKS / Lambda, validating a sampling configuration, or generating daemon deployment CLI and IaC templates. Triggers: X-Ray, distributed tracing, X-Ray daemon, sampling rules, service map, X-Ray SDK, OpenTelemetry, ADOT, Lambda Powertools, CloudWatch ServiceLens.
wellarchitected-workload-auditor
Audits AWS Well-Architected Tool workloads for review staleness, per-pillar high-risk issue counts, milestone tracking gaps, and remediation plan completeness. Emits a deterministic verdict (STALE_REVIEW | HIGH_RISK | CONFIG_GAP | OK) per workload with per-pillar risk breakdown, milestone trend analysis, and specific remediation. Use when reviewing Well-Architected workloads, checking WA review freshness, auditing high-risk issues per pillar, validating milestone coverage, or tracking remediation plans.
wellarchitected-review-operator
Operates AWS Well-Architected Tool reviews end-to-end — creates workloads, runs pillar reviews (operational excellence, security, reliability, performance, cost optimization, sustainability, Prosperity), answers questions with risk-tier rationale, generates reports and improvement plans, creates milestones, and integrates with Trusted Advisor and Well-Architected Labs. Runs deterministic pre-checks (workload existence, lens availability, pillar coverage), emits the exact update-answer / get-consolidated-report CLI behind a CONFIRM gate, and verifies risk-tier transitions. Emits READY | BLOCKED | COMPLETED. Use when creating a workload, answering pillar questions, generating an improvement plan, creating a milestone, or integrating with Trusted Advisor. Triggers: well-architected tool, well architected review, workload review, pillar, operational excellence, security, reliability, performance, cost optimization, sustainability, prosperity, improvement plan, milestone, trusted advisor.
wafv2-web-acl-deployer
Provisions AWS WAFv2 Web ACLs with correct production defaults: scope selection (CLOUDFRONT vs REGIONAL), managed rule groups (AWSManagedRulesCommonRuleSet, SQLiRuleSet, LinuxRuleSet, WindowsRuleSet, IPReputationList, BotControlRuleSet, AmazonIpReputationList, ATP), custom rules (byte match, IP set, regex pattern, geo match, size constraint, rate-based), rule priority ordering, logging destinations (Kinesis Firehose, CloudWatch Logs, S3), resource association (ALB, API Gateway, CloudFront), rate-based rules with aggregate keys (IP, forwarded IP), Challenge and CAPTCHA actions, and account takeover prevention. Emits a READY_TO_DEPLOY checklist. Use when creating a Web ACL, attaching managed rules, configuring rate limiting, enabling WAF logging, associating a Web ACL with an ALB/API Gateway/CloudFront, or configuring CAPTCHA/Challenge actions. Triggers: create WAF Web ACL, WAFv2, managed rule groups, WAF rate limiting, WAF logging, WAF CAPTCHA, WAF Challenge action, bot control, ATP managed rules.
wafv2-web-acl-auditor
Audits AWS WAFv2 Web ACL configurations to determine whether the ACL provides effective protection: default-action posture (Allow vs Block), managed-rule-group coverage gaps, rule effectiveness (BLOCK vs COUNT, shadow/bypass rules, stale exclusions), rate-based rule correctness, logging and visibility configuration, and text-transformation bypass vectors. Emits a deterministic verdict per Web ACL. Use when reviewing a WAFv2 Web ACL for security posture, checking managed-rule coverage, validating rate-limit configuration, auditing WAF logging, or hardening a WAF before production deployment. Triggers: WAF, WAFv2, Web ACL, managed rule group, AWSManagedRules, rate-based rule, CAPTCHA, Challenge, OverrideAction, Count mode, forwarded IP, Bot Control, ATP, text transformation, WAF logging, WAF visibility, OWASP, SQLi, XSS, RFI.
waf-rule-deployer
Provisions AWS WAFv2 rule sets with production defaults: Web ACL creation (CloudFront scope in us-east-1 vs Regional scope), rule groups (managed vs custom) with capacity planning in WCU (Web ACL Capacity Units, 1500 default per ACL), managed rule groups (AWSManagedRulesCommonRuleSet, AWSManagedRulesSQLiRuleSet, AWSManagedRulesAmazonIpReputationList, AWSManagedRulesLinuxRuleSet, AWSManagedRulesWindowsRuleSet, AWSManagedRulesUnixRuleSet), custom rules (byte-match, regex pattern set, geo-match, rate-based), IP set and regex pattern set resources, rule priority ordering (lowest number evaluated first), action types (allow, block, count, CAPTCHA, Challenge), bot control managed rule group, account takeover. Triggers: create WAF rule, WAF managed rule group, WAF custom rule, WAF byte match, WAF geo match, WAF rate based rule, WAF IP set, WAF regex pattern set, WAF WCU budget, WAF CloudFront ACL, WAF logging Kinesis Firehose, WAF bot control, WAF ATP, WAF label matching, WAF Challenge action, WAF CAPTCHA.