engyossefyossry-crypto/n8n-sentinel-circuit-breaker-or-resilient-webhook-queue-engine

This project is a ready made workflow for n8n, a visual automation tool where you build processes by connecting boxes on a canvas instead of writing a full program by hand. It solves a specific reliability problem: what happens when a webhook, meaning an incoming request that triggers your automation, keeps trying to reach a backend service that has gone down. Without protection, an automation like this can keep hammering a broken server over and over, making the outage worse. The workflow implements a pattern from traditional backend engineering called a circuit breaker, built entirely with native n8n nodes rather than any AI or language model components. It works in three states. In the normal, closed state, requests pass straight through to the target server as usual. If the target fails five times in a row, the circuit trips to an open state, where new requests are immediately redirected to a separate holding area called a dead letter queue instead of being sent to the already struggling server, protecting it from more load. A background check runs every two minutes while the circuit is open, and once that check succeeds, the workflow moves to a half open state where it lets exactly one test request through. If that test succeeds, the circuit fully resets to normal, if it fails, the workflow immediately trips back to the open state rather than flooding the server again. To avoid many failed requests retrying at the exact same moment and overwhelming a recovering server, the workflow calculates wait times using a growing delay that also includes some randomness, a common technique called exponential backoff with jitter. The current state and retry counts are stored using n8n's built in shared memory feature, so the circuit breaker remembers what state it is in across separate, unrelated runs of the workflow rather than starting fresh each time. Setting it up means importing a single workflow.json file into an existing n8n installation, then pointing two nodes at your own backend, one for the main API you are calling and one for a health check endpoint used by the background recovery test. The README also walks through a manual test plan using a free mock API service so you can watch the circuit breaker open, redirect traffic, and recover without needing a real backend to fail on purpose. The project is released under the MIT license.