Redis::CannotConnectError in CI

The Redis client could not establish a connection. The Redis service may not be up yet, or REDIS_URL points at the wrong host/port. Connecting to a starting cache/queue service is a common transient failure.

What this error means

Tests or boot fail with Redis::CannotConnectError (connection refused or timeout). Often intermittent while the Redis service container is still starting.

rails

Redis::CannotConnectError: Error connecting to Redis on
localhost:6379 (Errno::ECONNREFUSED)

Common causes

Service not ready

The app connected before the Redis service container began accepting connections.

Wrong URL/host/port

REDIS_URL or the configured host/port does not match the CI Redis service.

Redis not provisioned

The workflow never started a Redis service, but the test setup (Sidekiq, cache, ActionCable) needs one.

How to fix it

Provision Redis and wait for it

Start a Redis service, point REDIS_URL at it, and gate on readiness.

Terminal

until redis-cli -h localhost -p 6379 ping | grep -q PONG; do sleep 1; done
# REDIS_URL: redis://localhost:6379/0

Configure the client from env

Set REDIS_URL in the job env to the service address.
Add a redis service (or container) to the workflow.
Gate the test step on a successful PING.

How to prevent it

Add a Redis readiness gate before tests.
Drive Redis config from REDIS_URL matching the service.
On self-healing managed runners (Latchkey), transient Redis connection failures during service startup are auto-retried rather than failing the job.