Describe the bug

When Redis connection is lost or becomes unstable (ECONNRESET errors), Flowise enters a state where:

• UI appears functional and responsive
• Message requests are accepted but never respond
• System eventually returns 504 Gateway Timeout
• No automatic recovery mechanism exists
• Entire system becomes unusable until manual restart

Root Cause: buildChatflow.ts:986 contains await job.waitUntilFinished(queueEvents) with no timeout, causing indefinite blocking when Redis queue jobs cannot be processed.

Architecture Impact: Redis becomes a single point of failure that can render the entire system non-functional, even though core AI processing could work fine without queuing.

To Reproduce

1. Deploy Flowise in queue mode (MODE=queue) with Redis backend (e.g., Google Cloud Memorystore)
2. Configure with unstable TLS setup (e.g., NODE_TLS_REJECT_UNAUTHORIZED=0) or network conditions causing Redis ECONNRESET
3. Send a message through the UI
4. Observe Redis connection drops/resets in logs
5. Message request hangs indefinitely
6. All subsequent message requests also hang
7. System requires restart to recover

Expected behavior

Resilient behavior should include:

• Job processing timeout configuration (e.g., JOB_TIMEOUT environment variable)
• Graceful fallback to direct processing when Redis is unavailable
• Circuit breaker pattern for Redis connection failures
• Automatic retry logic with exponential backoff
• Health checks that detect Redis connectivity issues
• Error responses instead of indefinite hanging

Screenshots

No response

Flow

No response

Use Method

Docker

Flowise Version

3.0.4

Operating System

Linux

Browser

None

Additional context

Code Location: /packages/server/src/utils/buildChatflow.ts:986
// Currently: No timeout specified
const result = await job.waitUntilFinished(queueEvents)

// Should be: Configurable timeout
const result = await job.waitUntilFinished(queueEvents, {
timeout: parseInt(process.env.JOB_TIMEOUT || '300000') // 5 min default
})

Missing Resilience Patterns:

• No timeout on job.waitUntilFinished()
• No fallback when Redis unavailable
• No circuit breaker for Redis failures
• Health checks don't verify Redis connectivity
• No job processing timeout configuration via environment variables

BullMQ Support: BullMQ v5.45.2 supports timeouts, but Flowise doesn't expose this configuration.