SkillVaultskills Browse all 500 skills

Reliability · Version 1.0.0 · Reviewed 2026-08-02

Production Load Testing Advisor

Reduce production risk in traffic isolation and data safety with evidence, explicit trade-offs, and a verification plan.

4 method steps 4 documented failure modes 4 diagnostic checks 7 quality gates

Runs load tests against production safely with traffic tagging, isolation, and abort criteria.

₹99 one-time

Get this skill archive

What this skill helps you do

  • Traffic isolation
  • Data safety
  • Abort criteria

How Production Load Testing Advisor works

You provide

Impact window, telemetry, and dependency state

It inspects

Saturation and blast radius behind traffic isolation

It decides

A data safety plan that stabilizes before diagnosing

You verify

Detect, mitigate, and resolve times recorded separately

What it checks first

Production Load Testing Advisor runs load tests against production safely with traffic tagging, isolation, and abort criteria. Use it when the work involves Traffic isolation, Data safety, Abort criteria.

  1. User-visible impact and error-budget consumption rather than component health.
  2. Saturation signals — queue depth, pool utilization, connection counts — near the onset.
  3. Whether the system recovered on its own, which indicates saturation rather than corruption.
  4. The blast radius and what boundary should have contained it.

Failure modes it recognizes

  • Retry amplification turning a partial failure into a total outage.
  • A shared dependency creating correlated failure across supposedly independent services.
  • Slow resource exhaustion invisible until a hard limit is crossed.
  • A rollback blocked by an incompatible migration.

Answers it will reject

  • Treating the trigger as the root cause, which stops the analysis before the fragility is identified.
  • Adding a runbook step where a boundary would remove the failure mode.
  • Measuring availability as a mean, which hides regional and tenant-level outages.

Decision rules it applies

  • Stabilize user impact before completing diagnosis.
  • Bound every retry with a budget, jitter, and a circuit breaker.
  • Prefer removing a failure mode over detecting it faster.

Evidence it asks for

  • Record time-to-detect, time-to-mitigate, and time-to-resolve separately.
  • Quantify impact in customer terms: failed requests, affected accounts, duration.
  • Verify recovery with the same signal that detected the failure.

The method inside

  1. Map the artifact, actors, boundaries, and invariants relevant to traffic isolation.
  2. Trace concrete failure or abuse paths for data safety; do not report checklist items without a mechanism.
  3. Prioritize abort criteria findings by impact, likelihood, confidence, and cost of correction.
  4. Recommend the smallest defensible change, then define how an independent reviewer can verify it.

Deliverables

  • Traffic isolation assessment
  • Data safety decision and action plan
  • Abort criteria verification checklist

Evidence requirements

  • User-visible symptoms and SLO impact
  • Timeline, telemetry, deploys, and dependency state
  • Current mitigations and operational constraints

Quality gates

  • Every material claim traces to supplied evidence or is labeled as a hypothesis.
  • The response follows the declared deliverable contract.
  • No execution, access, measurement, or verification is invented.
  • Secrets and personal data are redacted rather than repeated.
  • The user receives a concrete independent verification step.
  • The relevant failure modes in this domain were considered rather than only the reported symptom.
  • No listed anti-pattern was recommended as a solution.

Example task

Input

We want to load test in production because staging never predicts real behavior. How do we do it safely?

Expected output

The instinct is right and the risk is contaminating real data. Tag synthetic traffic end to end so it can be excluded from analytics and billing, route it to real infrastructure but sandboxed data, and define an abort trigger on customer-facing latency...

Boundaries and compatibility

Ideal for

  • Traffic isolation: produce a decision or artifact grounded in supplied evidence.
  • Data safety: produce a decision or artifact grounded in supplied evidence.
  • Abort criteria: produce a decision or artifact grounded in supplied evidence.

Out of scope

  • Replacing incident command authority
  • Calling a trigger the root cause without a causal chain

Agent compatibility

  • GitHub Copilot custom agents
  • Claude Agent Skills / SKILL.md
  • Any instruction-following chat model

Tool policy: Advisory by default. No tools are assumed. If the host provides tools, use read-only evidence gathering unless the user explicitly approves a scoped write or execution action.