Chaos Engineering¶
Basiri, A., Behnam, N., de Rooij, R., Hochstein, L., Kosewski, L., Reynolds, J., & Rosenthal, C. (2016). Chaos Engineering. IEEE, 33(3), 35-41.
Cited by¶
20 citations across 20 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Antifragility
- Engineering and software: Chaos engineering deliberately injects failures into production systems so that recovery machinery is exercised and hardened, an empirical discipline Basiri and colleagues (2016) describe in their account of Netflix's controlled failure-injection practice; systems that are continuously perturbed develop deeper fault tolerance than systems kept artificially quiet.
This sourceDescribes Netflix's controlled failure-injection discipline for distributed systems — deliberately perturbing production to exercise and harden recovery machinery so continuously perturbed systems develop deeper fault tolerance than systems kept artificially quiet.
Supported in partVerified against the work's full text
Confirms Netflix injects failures into production to improve reliability; the comparative claim that perturbed systems develop deeper fault tolerance than quiet ones is not tested or stated.
“Chaos Monkey proved successful, as today inside of Netflix all engineers design their services to handle instance failures as a matter of course. That success encouraged us to extend the approach of injecting failures into the production system in order to improve reliability.”
- Engineering and software: Chaos engineering deliberately injects failures into production systems so that recovery machinery is exercised and hardened, an empirical discipline Basiri and colleagues (2016) describe in their account of Netflix's controlled failure-injection practice; systems that are continuously perturbed develop deeper fault tolerance than systems kept artificially quiet.
- Fuzzing
- In resilience engineering it is chaos-engineering practice — randomly killing instances or corrupting links to surface latent gaps.
This sourceFormalizes chaos engineering — injecting random failures (killing instances, corrupting links) into production to surface latent gaps, the resilience-substrate instance of fuzzing.
SupportedVerified against the work's full text
Describes Chaos Monkey randomly terminating production instances, experiments that sever network connections, and experiments ending when a weakness is uncovered.
“Netflix has been running an internal service called Chaos Monkey [2], which randomly selects virtual machine instances that host our production services and terminates them.”
- In resilience engineering it is chaos-engineering practice — randomly killing instances or corrupting links to surface latent gaps.
- Inoculation Theory
- Reliability engineering. Game days, fault injection, and fuzz testing — pre-attacking one's own system in graded doses to surface weakness while consequences remain bounded.
This sourceDescribes deliberately injecting controlled faults (game days, fault injection) to build resilience to rare production failures.
Supported in partVerified against the work's full text
Supports deliberately injecting faults into one's own system to uncover weakness, with scope reduced to a subset of users; it does not mention game days, fuzz testing or graded doses.
“At the conclusion of an experiment, we either have more confidence in the system’s ability to maintain behavior in the presence of our variables, or a weakness has been uncovered”
- Reliability engineering. Game days, fault injection, and fuzz testing — pre-attacking one's own system in graded doses to surface weakness while consequences remain bounded.
- Rehearsal
- In engineering operations, incident-response game days, chaos exercises, and disaster-recovery failovers maintain response capability before it is needed under real load.
This sourceDefines chaos engineering / game days: deliberate fault injection into production-like systems to rehearse the whole incident response under near-live conditions.
Supported in partVerified against the work's full text
The paper describes Netflix chaos exercises and region-failover tests in production to improve reliability, but not game days, incident-response practice, or disaster-recovery drills.
“we perform "Chaos Kong" exercises that simulate the failure of an entire Amazon EC2 region”
- In engineering operations, incident-response game days, chaos exercises, and disaster-recovery failovers maintain response capability before it is needed under real load.
- Robustness
- To handle this envelope, the platform implements multiple independent robustness mechanisms: client libraries implement retries with jitter to handle transient failures; API gateways implement rate limiting and backpressure with explicit prioritization of critical transactions over lower-value operations (margin by prioritization); each service runs with independent capacity headroom above nominal peak load (margin by overprovisioning); the fraud-detection subsystem has a fail-safe default (decline on service failure) preserving safety at availability cost; the entire system has been tested under simulated peak loads (100x+ baseline) and induced component failures (chaos engineering)
This sourceChaos engineering stress testing replica failures.
Supported in partVerified against the work's full text
Basiri et al. support graceful-degradation testing and measured validation of a production system, but not the claim's specified-envelope and independent-mechanism framing.
“we also run Failure Injection Testing (FIT) exercises where we cause requests between Netflix services to fail and verify that the system degrades gracefully.”
- To handle this envelope, the platform implements multiple independent robustness mechanisms: client libraries implement retries with jitter to handle transient failures; API gateways implement rate limiting and backpressure with explicit prioritization of critical transactions over lower-value operations (margin by prioritization); each service runs with independent capacity headroom above nominal peak load (margin by overprovisioning); the fraud-detection subsystem has a fail-safe default (decline on service failure) preserving safety at availability cost; the entire system has been tested under simulated peak loads (100x+ baseline) and induced component failures (chaos engineering)
- Stability
- Structural engineering: a building's return to its design configuration under wind, seismic, or live load; buckling analysis is the search for loss of stability under increasing load. Neural circuits and software: firing rates settling at characteristic patterns via excitation-inhibition balance; runtime recovery from transient errors via retries, circuit breakers, and autoscaling, with chaos engineering as deliberate perturbation to test stability.
This sourceDescribes deliberate perturbation of distributed systems to test runtime stability and recovery.
Supported in partVerified against the work's full text
Supports chaos engineering as deliberate failure injection to test a distributed system's resilience; the source does not mention retries, circuit breakers, autoscaling or neural circuits.
“Chaos Engineering is the discipline of experimenting on a distributed system in order to build confidence in its”
- Structural engineering: a building's return to its design configuration under wind, seismic, or live load; buckling analysis is the search for loss of stability under increasing load. Neural circuits and software: firing rates settling at characteristic patterns via excitation-inhibition balance; runtime recovery from transient errors via retries, circuit breakers, and autoscaling, with chaos engineering as deliberate perturbation to test stability.
- Stability-Induced Fragility
- The cleanest modern transfer runs from finance and ecology into software operations: chaos engineering is the stability-induced-fragility counter-intervention applied to distributed systems, deliberately re-introducing the controlled failure that the calm would otherwise let the system forget how to survive.
This sourceChaos engineering as the controlled-stressor counter-intervention deliberately reintroducing failure into distributed systems.
Supported in partVerified against the work's full text
Supports chaos engineering as deliberately injecting failure into distributed systems; the finance/ecology origin and the idea that calm erodes survival skills are not in the source.
“Chaos Engineering is the discipline of experimenting on a distributed system in order to build confidence in its”
- The cleanest modern transfer runs from finance and ecology into software operations: chaos engineering is the stability-induced-fragility counter-intervention applied to distributed systems, deliberately re-introducing the controlled failure that the calm would otherwise let the system forget how to survive.
- Wargaming
- Negotiation training: devil's-advocate role-play and rehearsal against a sparring partner who attacks the deal. Public-health preparedness: surge tabletops and pandemic exercises. Product strategy: pre-mortems with a competitor-in-the-room role and "kill the product" exercises. Legal preparation and software resilience: moot court and mock trial; chaos engineering and fault-injection campaigns where the chaos is the adaptive red.
This sourceDefines chaos engineering / fault-injection as deliberately injecting failure into production systems to surface hidden assumptions about resilience.
Supported in partVerified against the work's full text
Basiri et al. define chaos engineering as deliberately injecting failures into production systems to test resilience, which covers only the software-resilience clause of this multi-domain claim.
“Chaos Engineering is the discipline of experimenting on a distributed system in order to build confidence in its capability to withstand turbulent conditions in production.”
- Negotiation training: devil's-advocate role-play and rehearsal against a sparring partner who attacks the deal. Public-health preparedness: surge tabletops and pandemic exercises. Product strategy: pre-mortems with a competitor-in-the-room role and "kill the product" exercises. Legal preparation and software resilience: moot court and mock trial; chaos engineering and fault-injection campaigns where the chaos is the adaptive red.
Mechanisms¶
- Ablation or Perturbation Test
- Accommodation Failure and Recovery Rehearsal
- This is the discipline of chaos engineering — deliberately injecting failures, including compound ones, to learn where a system actually breaks rather than where you assume it will.
This sourceDefines chaos engineering as deliberately injecting realistic failure conditions and observing actual system behavior to uncover weaknesses that assumptions and specifications miss.
- This is the discipline of chaos engineering — deliberately injecting failures, including compound ones, to learn where a system actually breaks rather than where you assume it will.
- Controlled Stress-Pulse Test
- Rather than wait for a real incident, they design a stress-pulse test in the spirit of chaos engineering.
This sourceDefines chaos engineering as deliberately running controlled failure experiments to expose weaknesses before real incidents do.
- Rather than wait for a real incident, they design a stress-pulse test in the spirit of chaos engineering.
- Failure-Injection Test
- Its software cousin is chaos engineering
This sourceDefines chaos engineering as experiments that inject failures into running production systems to test resilience.
- Its software cousin is chaos engineering
- Inoculation Refresh Drill
- So the team runs recurring GameDays: on a set cadence
This sourceDescribes recurring chaos exercises that deliberately inject failures into a limited slice of a production system.
- So the team runs recurring GameDays: on a set cadence
- Local Ablation or Lesion Probe
- The discipline is to bound the blast radius, run in a controlled window, and feed the resulting footprint to the canaries that will watch for it in the wild.
This sourceFrames chaos engineering as controlled experimentation whose potential impact should be minimized and contained.
- The discipline is to bound the blast radius, run in a controlled window, and feed the resulting footprint to the canaries that will watch for it in the wild.
- Path Readiness Drill
- On the other days of the quarter nothing switched over; the drill's entire value was the tested confidence, and the two silent failures it caught.
This sourceDefines chaos engineering as empirical experimentation that builds confidence in a system’s capacity to withstand turbulent conditions.
- On the other days of the quarter nothing switched over; the drill's entire value was the tested confidence, and the two silent failures it caught.
- Perturbation Probe
- The team runs a Perturbation Probe in the tradition of chaos engineering: during a controlled window, they deliberately terminate a random service instance in the live cluster and inject network latency between two internal services, at magnitudes calibrated to real-world incidents rather than to catastrophe.
This sourceDescribes controlled chaos-engineering experiments that inject realistic instance and network failures in production.
- The team runs a Perturbation Probe in the tradition of chaos engineering: during a controlled window, they deliberately terminate a random service instance in the live cluster and inject network latency between two internal services, at magnitudes calibrated to real-world incidents rather than to catastrophe.
- Perturbation-Response Map
- Stochastic Robustness Test
- So it runs a chaos-engineering test: a tool randomly terminates production service instances — but only within a bounded blast radius and only during business hours, with an always-available kill switch — to verify the system reroutes and self-heals without user-visible disruption.
This sourceDefines chaos engineering as controlled experimentation on distributed systems while minimizing the blast radius.
- So it runs a chaos-engineering test: a tool randomly terminates production service instances — but only within a bounded blast radius and only during business hours, with an always-available kill switch — to verify the system reroutes and self-heals without user-visible disruption.
- Substitution Drill
- Its strength is that it converts an assumed alternative into a proven one and calibrates the reserve and firebreak to reality — it is the only mechanism here that can catch "the second source cannot actually take the load in time," which is precisely the failure that surfaces at the worst possible moment otherwise.
This sourceBasiri and colleagues describe controlled failure experiments as a way to replace assumption with empirical confidence about system behavior.
- Its strength is that it converts an assumed alternative into a proven one and calibrates the reserve and firebreak to reality — it is the only mechanism here that can catch "the second source cannot actually take the load in time," which is precisely the failure that surfaces at the worst possible moment otherwise.
- Withdrawal Rebound Drill
- Its lineage is chaos engineering — the practice of deliberately injecting failure into a running system to prove its resilience empirically rather than trusting the design
This sourceDescribes chaos engineering as injecting failures into running systems to test resilience empirically.
- Its lineage is chaos engineering — the practice of deliberately injecting failure into a running system to prove its resilience empirically rather than trusting the design
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Read against the text for 8 of 20 citations: 1 supported, 7 supported in part. Each verdict is shown under its citation below, with what in the work backs the sentence.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:be815ae9129b · see in the full table