fix: add Redis connection resilience settings to prevent infinite hangs

[utsav-bx]

Problem

Fixes #5126 - Flowise hangs indefinitely when Redis connections drop, causing ECONNRESET errors during queue processing.

Solution

This PR implements a two-part fix to prevent indefinite hangs:

1. Redis Connection Resilience (Optional Configuration)

Adds configurable Redis connection settings via environment variables without hardcoded defaults:

• REDIS_CONNECT_TIMEOUT - Connection timeout in milliseconds
• REDIS_COMMAND_TIMEOUT - Command timeout in milliseconds
• REDIS_MAX_RETRIES_PER_REQUEST - Maximum number of retries per request
• REDIS_LAZY_CONNECT - Enable lazy connection (true/false)

These settings are only applied if explicitly set, allowing Redis to use its own defaults otherwise.

2. Job Execution Timeouts (Core Fix)

Adds JOB_TIMEOUT environment variable (default: 300000ms / 5 minutes) and implements explicit timeouts on all job.waitUntilFinished() calls across:

• buildChatflow.ts
• agentflowv2-generator/index.ts
• documentstore/index.ts (5 occurrences)
• nodes/index.ts
• upsertVector.ts

This prevents jobs from hanging indefinitely when Redis connections fail.

Key Changes

• Enhanced packages/server/src/queue/QueueManager.ts with conditional Redis resilience settings
• Updated packages/server/.env.example with new configuration options
• Added explicit timeouts to all 9 waitUntilFinished() calls
• Maintains full backward compatibility with existing Redis environment variables

Addresses PR Feedback

• ✅ Removed hardcoded defaults per @HenryHengZJ's feedback
• ✅ Uses conditional property spreading - only includes settings when env vars are defined
• ✅ Allows Redis to use its own internal defaults when not configured

Testing

• ✅ Build passes with pnpm build
• ✅ TypeScript compilation successful
• ✅ All existing Redis env vars preserved
• ✅ Backward compatible - works with or without new env vars