> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/resilience-testing/chaos-engineering/faults/chaos-fault-categories/cloud-foundry/cloud-foundry.md). # Chaos faults for Cloud Foundry Catalog of Cloud Foundry chaos faults that disrupt apps, JVM runtimes, and the network between app instances and their dependencies. ### Introduction Cloud Foundry chaos faults disrupt the running state of an app, the JVM inside an app instance, or the network that connects an app to its dependencies. Use them to validate platform self-healing (CF restarting crashed instances), application resilience (timeouts, retries, fallbacks), and downstream behavior (consumer error rates, alert fidelity) on TAS, PCF, or open-source Cloud Foundry deployments. All Cloud Foundry faults run from a Linux chaos infrastructure (LCI) that authenticates to the Cloud Foundry API and, for JVM and network faults, to the BOSH director that manages the Diego cells. Go to [Cloud Foundry chaos deployment](/resilience-testing/chaos-engineering/faults/chaos-fault-categories/cloud-foundry/cf-components-and-arch.md) to read the supported LCI deployment models, and [Cloud Foundry permissions](/resilience-testing/chaos-engineering/faults/chaos-fault-categories/cloud-foundry/permissions.md) to set up the required CF and BOSH credentials.
CF app stopStop a Cloud Foundry app for a configurable duration and then re-start it, so you can test how the platform and consumers react when the app goes offline.cf, cloud-foundry, app, stop#cf-app-stop
CF app container killKill the container of one or more app instances and let Cloud Foundry reschedule them so you can validate platform self-healing and peer absorption.cf, cloud-foundry, app, container, kill#cf-app-container-kill
CF app route unmapTemporarily unmap a route from a Cloud Foundry app so you can test how upstream gateways and consumers handle a route that returns 404 from the CF router.cf, cloud-foundry, app, route#cf-app-route-unmap
CF app JVM CPU stressDrive sustained CPU usage inside the JVM of a Java app instance for a configurable duration so you can test the app under CPU saturation.cf, cloud-foundry, app, jvm, cpu#cf-app-jvm-cpu-stress
CF app JVM memory stressDrive heap or non-heap memory pressure inside the JVM of a Java app instance so you can test GC behavior and out-of-memory handling.cf, cloud-foundry, app, jvm, memory#cf-app-jvm-memory-stress
CF app JVM trigger GCForce a full garbage-collection cycle inside the JVM of a Java app instance so you can measure pause time and tail-latency impact.cf, cloud-foundry, app, jvm, gc#cf-app-jvm-trigger-gc
CF app JVM method exceptionForce a specific JVM method to throw a configurable exception so you can validate error-handling paths in callers.cf, cloud-foundry, app, jvm, exception#cf-app-jvm-method-exception
CF app JVM method latencyAdd artificial latency to every invocation of a specific JVM method so you can simulate a slow downstream call at the method boundary.cf, cloud-foundry, app, jvm, latency#cf-app-jvm-method-latency
CF app JVM modify returnOverride the return value of a specific JVM method so you can test defensive checks and fallbacks against unexpected values.cf, cloud-foundry, app, jvm, return#cf-app-jvm-modify-return
CF app network latencyAdd latency on the egress traffic of an app instance so you can simulate a slow downstream dependency at the network layer.cf, cloud-foundry, app, network, latency#cf-app-network-latency
CF app network lossDrop a configurable percentage of egress packets from an app instance so you can test retransmissions, retries, and circuit-breaker behavior.cf, cloud-foundry, app, network, loss#cf-app-network-loss
CF app network corruptionCorrupt a configurable percentage of egress packets from an app instance so you can validate TCP retransmissions and protocol parsers.cf, cloud-foundry, app, network, corruption#cf-app-network-corruption
CF app network duplicationDuplicate a configurable percentage of egress packets from an app instance so you can validate idempotency and deduplication logic.cf, cloud-foundry, app, network, duplication#cf-app-network-duplication
#### CF app stop CF app stop stops a Cloud Foundry app for a configurable duration and then re-starts it. Use it to validate how the platform, routers, and downstream consumers behave when an app goes offline cleanly. Use cases * Validate consumer fallbacks and retries when the app returns 5xx from the CF router. * Confirm CF restarts the app and reports it healthy after `duration` elapses. * Tune alert thresholds around route-level health checks. View details #### CF app container kill CF app container kill terminates the container of one or more app instances and lets Cloud Foundry reschedule them. Use it to validate platform self-healing and peer absorption. Use cases * Confirm peer instances absorb traffic while the killed instance is rescheduled. * Validate Diego restarts the instance inside its expected window. * Verify in-flight requests fail cleanly so callers retry. View details #### CF app route unmap CF app route unmap detaches a specific route from an app for a configurable duration, then re-maps it. The app keeps running; only its inbound route is disrupted. Use cases * Validate gateway and consumer behavior when a route returns 404 from the CF router. * Confirm secondary routes mapped to the same app continue serving. * Practice runbooks for accidental route removal. View details #### CF app JVM CPU stress CF app JVM CPU stress drives high CPU usage inside the JVM of a Java app instance for a configurable duration. Use it to test the app and autoscaler under sustained CPU pressure. Use cases * Measure latency under CPU saturation. * Validate autoscaling rules trigger and pull traffic away from the stressed instance. * Surface thread-pool and concurrency bugs. View details #### CF app JVM memory stress CF app JVM memory stress drives sustained heap or non-heap memory pressure inside the JVM of a Java app instance. Use it to test GC behavior and `OutOfMemoryError` handling. Use cases * Surface long GC pauses under heap pressure. * Detect runaway metaspace consumption with non-heap pressure. * Validate memory-based autoscaling. View details #### CF app JVM trigger GC CF app JVM trigger GC forces a full garbage-collection cycle inside the JVM of a Java app instance. Use it to measure stop-the-world pause times and validate that health checks tolerate them. Use cases * Quantify worst-case GC pause time. * Confirm liveness/readiness probes do not falsely fail during a full GC. * Measure the tail-latency cost of a GC under production-shaped traffic. View details #### CF app JVM method exception CF app JVM method exception forces a specific JVM method to throw a configurable exception. Use it to validate error-handling paths, retry budgets, and circuit-breaker behavior. Use cases * Confirm catch blocks map the exception to the right user-visible response. * Validate retry and circuit-breaker thresholds. * Test observability tags for the configured exception class. View details #### CF app JVM method latency CF app JVM method latency adds artificial delay to every invocation of a specific JVM method. Use it to simulate a slow downstream call at the method boundary. Use cases * Tune timeouts to trip before the user-visible SLO is breached. * Quantify the latency contribution of a single slow method. * Validate caller retry budgets do not amplify the slowdown. View details #### CF app JVM modify return CF app JVM modify return overrides the return value of a specific JVM method. Use it to test defensive checks and fallbacks against unexpected values like `null` or wrong-type returns. Use cases * Validate null-safety in callers of a non-null method. * Force a feature flag off by overriding its accessor. * Simulate a poisoned cache layer. View details #### CF app network latency CF app network latency adds a configurable amount of latency (with optional jitter) on the egress traffic of an app instance. Use it to simulate a slow downstream dependency at the network layer. Use cases * Simulate a slow database or third-party API. * Quantify how added round-trip time affects user-visible P99. * Approximate cross-region latency to test caching decisions. View details #### CF app network loss CF app network loss drops a configurable percentage of egress packets from an app instance. Use it to test retransmissions, timeouts, and circuit-breaker behavior under packet loss. Use cases * Validate the app's retry budget on a flaky downstream. * Confirm circuit breakers open at the configured loss rate. * Test alert tuning for elevated TCP retransmissions. View details #### CF app network corruption CF app network corruption corrupts a configurable percentage of egress packets from an app instance. Corrupted packets are discarded by the receiver, triggering retransmissions. Use cases * Quantify retransmission overhead on end-to-end latency. * Test protocol parsers reject malformed frames cleanly. * Validate L4 monitoring detects elevated retransmission rates. View details #### CF app network duplication CF app network duplication duplicates a configurable percentage of egress packets from an app instance. Use it to validate idempotency assumptions and deduplication logic. Use cases * Confirm duplicate HTTP requests do not cause double writes. * Validate deduplication keys on message-bus consumers. * Test that UDP receivers tolerate duplicate datagrams. View details {% @harness-feedback/feedback %}