Istio Traffic Resilience Review
Evaluate route behavior and failure amplification against stated service objectives.
Decision and boundary
Does the proposed traffic behavior avoid unacceptable routing and failure amplification?
Operate as static-review. Read supplied files only; never collect live evidence, probe, mutate, or inherit ambient credentials. Host-enforced tool permissions remain mandatory.
Required inputs
Require intended behavior/access, rendered resources, exact environment/version/mode scope, and an evidence ledger with capture times and visibility limits. Continue bounded analysis when fields are missing, but mark affected conclusions needs-review.
Workflow
- Record user-facing deadline, critical operations, idempotency guarantees, load/capacity constraints and intended canary behavior.
- Resolve API family, mode, attachment, route precedence, hosts, destinations, ports and versioned backends; delegate Gateway attachment to its owner.
- Interpret relative route weights separately from percentages and evaluate destination health/availability evidence without inventing runtime distribution.
- Inventory retry owners across caller, mesh and dependencies; compare total/per-try deadlines, error eligibility, connection pools, ejection and remaining capacity.
- Treat mirroring and fault injection as business-affecting changes. Require isolation, data-handling scope, explicit experiment approval and stop criteria.
- Return route and resilience findings with an abort/rollback plan and paired tests; do not execute them.
Reference loading
Read evidence-boundary.md, applicability.md, and output-contract.md first. Load the following only for the active branch.
| Reference |
Load when |
| routing-and-weights.md |
Inspect route selection, destinations and relative weights. |
| timeouts-retries-and-capacity.md |
Review retry budgets, deadlines and overloaded dependencies. |
| mirroring-and-faults.md |
Review any real-traffic duplication or fault experiment. |
| negative-tests.md |
Define resilience and routing counterexamples. |
| official-sources.md |
Verify a version-sensitive rule or source disagreement. |
| review-contract.json |
Produce or validate a portable structured review. |
Deliverable
Route table, resilience budget, canary criteria, side-effect boundaries, and rollback proposal.
Separate observed evidence, derived conclusions, assumptions and unknowns. Include one counterexample for every consequential finding. Never describe unperformed tests as passing or a bounded review as production certification.
1---2name: istio-traffic-resilience-review3description: Review Istio VirtualService and DestinationRule or supplied Gateway API4 traffic intent for routing, weights, timeouts, retries, outlier detection, circuit-breaking,5 mirroring and fault injection. Use for canaries, retry storms, unexplained routing,6 write-side-effect risks or resilience changes. Evaluate target version and mesh7 mode before applying a rule. Do not inject faults, mirror traffic, probe services8 or change live configuration.9allowed-tools: Read Grep Glob10metadata:11 source: https://skillmd.com/skills/vincentchuwaichow/istio-traffic-resilience-review12 author: 'github: VincentChuWaiChow'13 version: 0.1.014 category: resilience15 lifecycle: beta16 execution_tier: static-review17---1819# Istio Traffic Resilience Review20Evaluate route behavior and failure amplification against stated service objectives.21## Decision and boundary22Does the proposed traffic behavior avoid unacceptable routing and failure amplification?23Operate as static-review. Read supplied files only; never collect live evidence, probe, mutate, or inherit ambient credentials. Host-enforced tool permissions remain mandatory.24## Required inputs25Require intended behavior/access, rendered resources, exact environment/version/mode scope, and an evidence ledger with capture times and visibility limits. Continue bounded analysis when fields are missing, but mark affected conclusions `needs-review`.26## Workflow271. Record user-facing deadline, critical operations, idempotency guarantees, load/capacity constraints and intended canary behavior.282. Resolve API family, mode, attachment, route precedence, hosts, destinations, ports and versioned backends; delegate Gateway attachment to its owner.293. Interpret relative route weights separately from percentages and evaluate destination health/availability evidence without inventing runtime distribution.304. Inventory retry owners across caller, mesh and dependencies; compare total/per-try deadlines, error eligibility, connection pools, ejection and remaining capacity.315. Treat mirroring and fault injection as business-affecting changes. Require isolation, data-handling scope, explicit experiment approval and stop criteria.326. Return route and resilience findings with an abort/rollback plan and paired tests; do not execute them.3334## Reference loading35Read [evidence-boundary.md](references/evidence-boundary.md), [applicability.md](references/applicability.md), and [output-contract.md](references/output-contract.md) first. Load the following only for the active branch.36| Reference | Load when |37| --- | --- |38| [routing-and-weights.md](references/routing-and-weights.md) | Inspect route selection, destinations and relative weights. |39| [timeouts-retries-and-capacity.md](references/timeouts-retries-and-capacity.md) | Review retry budgets, deadlines and overloaded dependencies. |40| [mirroring-and-faults.md](references/mirroring-and-faults.md) | Review any real-traffic duplication or fault experiment. |41| [negative-tests.md](references/negative-tests.md) | Define resilience and routing counterexamples. |42| [official-sources.md](references/official-sources.md) | Verify a version-sensitive rule or source disagreement. |43| [review-contract.json](references/review-contract.json) | Produce or validate a portable structured review. |4445## Deliverable46Route table, resilience budget, canary criteria, side-effect boundaries, and rollback proposal.47Separate observed evidence, derived conclusions, assumptions and unknowns. Include one counterexample for every consequential finding. Never describe unperformed tests as passing or a bounded review as production certification.