Kubernetes Workload Design
Overview
Use this skill to design Kubernetes workloads that scale predictably and roll out safely under real traffic behavior.
Scope Boundaries
- Use this skill when the task matches the trigger condition described in
description. - Do not use this skill when the primary task falls outside this skill's domain.
Shared References
- Autoscaling and rollout decision rules:
references/autoscaling-and-rollout-decision-rules.md
Templates And Assets
- Workload sizing template:
assets/workload-sizing-template.md
- Rollout strategy checklist:
assets/rollout-strategy-checklist.md
Inputs To Gather
- Traffic profile and latency/SLO targets.
- CPU/memory/concurrency characteristics.
- Rollout risk tolerance and availability requirements.
- Observability signals for scaling and rollback decisions.
Deliverables
- Workload sizing and scaling plan.
- Rollout strategy with guardrails and rollback triggers.
- Resilience assumptions and saturation behavior notes.
- Verification plan for load and deployment behavior.
Workflow
- Define resource and scaling assumptions in
assets/workload-sizing-template.md. - Choose scaling/rollout strategy using
references/autoscaling-and-rollout-decision-rules.md. - Validate rollout readiness via
assets/rollout-strategy-checklist.md. - Run representative load and rollout verification.
- Publish residual capacity and rollout risks with owners.
Quality Standard
- Resource sizing reflects measured workload behavior.
- Autoscaling avoids oscillation and delayed recovery.
- Rollout controls match service criticality.
- Rollback criteria are objective and monitored.
Failure Conditions
- Stop when workload design lacks safe rollout or capacity guarantees.
- Stop when autoscaling signals do not correlate with user impact.
- Escalate when saturation risk remains unmitigated.
Source: KentoShimizu/sw-agent-skills — distributed by TomeVault.