Netflix/Hystrix

Features

• Circuit Breakers - Implements circuit breakers that stop traffic to failing services to prevent cascading failures in distributed systems.
• Request Batching - Groups multiple concurrent calls into a single batch execution to reduce the total load on downstream systems.
• Dependency-Specific Thread Pools - Assigns each remote dependency to a dedicated thread pool to prevent latency in one service from saturating the system.
• Semaphore-Based Isolation - Assigns commands to dedicated thread pools or semaphores to isolate dependency latency.
• Dependency-Isolated - Uses dedicated thread pools to prevent latency in one dependency from saturating the entire system.
• Cascading Failure Preventions - Prevents cascading failures by isolating remote calls and providing fallback logic when downstream services fail.
• Semaphore-Based Command Isolation - Limits concurrent execution of non-network commands using semaphores to avoid thread management overhead.
• Fault-Tolerant Fallbacks - Returns a default or stubbed response when a primary command fails, times out, or is rejected.
• Service Call Fallbacks - Provides logic that returns default responses or alternative behaviors when a remote service call fails.
• Fault Tolerance - Wraps potentially failing remote calls in command objects to isolate latency and provide fallback logic.
• Fault Tolerance Implementation - Uses thread pools and semaphores to isolate dependencies, ensuring latency in one service does not saturate the system.
• Command-Based Fault Isolation - Wraps potentially failing remote calls in command objects that isolate latency and handle timeouts gracefully.
• Graceful Degradation - Implements architectural patterns for ensuring system stability through graceful degradation and fallback logic during partial failures.
• Request Batching - Combines multiple individual requests into a single execution to reduce the total number of remote calls.
• Circuit Breaking States - Manages service availability states like closed, open, and half-open based on failure rates to isolate failing services.
• Cascading Failure Prevention - Prevents cascading failures using circuit breakers and isolation mechanisms to ensure graceful degradation when external services fail.
• Collapsing Batch Requests - Runs multiple remote calls concurrently while caching identical requests and collapsing them into a single batch.
• Request-Scoped Caching - Stores remote call results per request context to prevent redundant network calls within a single request.
• Parallel Provider Calls - Invokes multiple service providers simultaneously to reduce total response latency and improve throughput.
• Service Health Monitoring - Tracks and reports on the operational status and performance metrics of internal application services.
• Asynchronous Command Execution - Queues dependency calls for asynchronous execution, returning a future to retrieve results without blocking the calling thread.
• Non-Blocking Execution - Implements non-blocking execution patterns to handle long tasks without freezing the main execution thread.
• Request Collapsing - Ensures only one concurrent request fetches data for a missing key while others await the result.
• Concurrency Optimizers - Optimizes throughput by executing requests in parallel and collapsing concurrent calls into single batches.
• Synchronous Command Execution - Executes commands and blocks the calling thread until a result or exception is returned.
• Metric Streaming - Models performance data as a real-time stream for off-box aggregation without blocking execution threads.
• Real-Time Monitoring Dashboards - Streams latency, error rates, and circuit states to real-time dashboards to reflect system health.
• Service Health Metrics Streaming - Streams real-time latency and error metrics to dashboards for immediate visibility into distributed system health.
• Service Health Monitoring - Tracks operational status and health of internal services through streaming latency and error metrics.
• Command Result Observables - Exposes the results of wrapped remote calls as observables, allowing non-blocking asynchronous reactions to emitted values.
• Reactive Stream Libraries - Wraps remote calls as asynchronous data streams to enable non-blocking reactions to values and errors.
• Distributed Applications - Library for latency and fault tolerance in remote systems.

Star history