Name: scalability-and-reliability
Author: joshua-noel

Imported from joshua-noel/skills (skills/system-design/scalability-and-reliability/SKILL.md). Install upstream with npx skills add joshua-noel/skills --skill scalability-and-reliability. Copyright stays with the author (MIT).

Scalability And Reliability

Trigger

Reliability incidents expose unclear limits.
A design needs SLO-driven review.

Inputs

Traffic and data forecasts
Critical journeys
SLOs/RTO/RPO
Dependency limits and failure history

Procedure

Define service-level indicators and realistic objectives per journey.
Model steady, peak, burst, and pathological load.
Find bottlenecks and finite resources across the full path.
Select partitioning, caching, queuing, batching, backpressure, and admission control only where needed.
Design timeouts, retries with jitter, circuit breaking, idempotency, and bulkheads coherently.
Specify degradation modes and customer-visible behavior.
Plan backup, failover, restore, and disaster exercises.
Create load, chaos, and recovery validation plans.

Output contract

Capacity model
Reliability architecture
Failure-mode table
SLO and validation plan

Quality gates

Retry budgets do not amplify outages.
Queues have bounds and poison-message handling.
Failover is tested, not assumed.
SLOs map to user journeys and alerting.

Handoffs and dependencies

performance-optimization
operations/incident-operations
operations/deployment-readiness
security/threat-modeling

Boundaries

Do not promise “infinite scale.”
Do not add distributed components without identifying the bottleneck they solve.
Do not confuse redundancy with recoverability.

{ "ocm": "1", "id": "joshua-noel-skills-scalability-and-reliability", "kind": "skill", "name": "scalability-and-reliability", "description": "Design capacity, resilience, graceful degradation, and recovery around concrete load and failure scenarios. Use when a system must handle growth, bursts, regional failure, or strict availability targets.", "publisher": "joshua-noel", "version": "1.0.0", "capabilities": { "domains": [ "general" ], "tags": [ "skill-md", "github" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "Design capacity, resilience, graceful degradation, and recovery around concrete load and failure scenarios. Use when a system must handle growth, bursts, regional failure, or strict availability targets." ], "primary": false, "metadata": { "source": { "provider": "github", "repository": "https://github.com/joshua-noel/skills", "path": "skills/system-design/scalability-and-reliability/SKILL.md", "ref": "86190a19d31a53a39e526e9ae6bc6b4e3fab79ce", "url": "https://github.com/joshua-noel/skills/blob/86190a19d31a53a39e526e9ae6bc6b4e3fab79ce/skills/system-design/scalability-and-reliability/SKILL.md", "key": "joshua-noel/skills/skills/system-design/scalability-and-reliability/SKILL.md" }, "license": "MIT" }, "instructions": "# Scalability And Reliability\n\n## Trigger\n\n- Reliability incidents expose unclear limits.\n- A design needs SLO-driven review.\n\n## Inputs\n\n- Traffic and data forecasts\n- Critical journeys\n- SLOs/RTO/RPO\n- Dependency limits and failure history\n\n## Procedure\n\n1. Define service-level indicators and realistic objectives per journey.\n2. Model steady, peak, burst, and pathological load.\n3. Find bottlenecks and finite resources across the full path.\n4. Select partitioning, caching, queuing, batching, backpressure, and admission control only where needed.\n5. Design timeouts, retries with jitter, circui", "cost": { "context_tokens": 366 } }

scalability-and-reliability

About