Redis connection lost and command aborted. code: 'UNCERTAIN_STATE'

We have a node.js app deployed on Heroku and we are using kue to run some background jobs. It usually works fine, but few days back we saw following errors in our production server:

```
Jul 10 11:19:23 refocus app/web.6: Error removing 4628051 { AbortError: Redis connection lost and command aborted. It might have been processed.
Jul 10 11:19:23 refocus app/web.6: at RedisClient.flush_and_error (/app/node_modules/kue/node_modules/redis/index.js:357:23)
Jul 10 11:19:23 refocus app/web.6: at RedisClient.connection_gone (/app/node_modules/kue/node_modules/redis/index.js:659:14)
Jul 10 11:19:23 refocus app/web.6: at Socket. (/app/node_modules/kue/node_modules/redis/index.js:293:14)
Jul 10 11:19:23 refocus app/web.6: at Socket.g (events.js:286:16)
Jul 10 11:19:23 refocus app/web.6: at emitNone (events.js:91:20)
Jul 10 11:19:23 refocus app/web.6: at Socket.emit (events.js:185:7)
Jul 10 11:19:23 refocus app/web.6: at endReadableNT (_stream_readable.js:975:12)
Jul 10 11:19:23 refocus app/web.6: at _combinedTickCallback (internal/process/next_tick.js:74:11)
Jul 10 11:19:23 refocus app/web.6: at process._tickDomainCallback [as _tickCallback] (internal/process/next_tick.js:122:9)
Jul 10 11:19:23 refocus app/web.6: code: 'UNCERTAIN_STATE',
Jul 10 11:19:23 refocus app/web.6: command: 'EXEC',
Jul 10 11:19:23 refocus app/web.6: errors:
Jul 10 11:19:23 refocus app/web.6: [ { AbortError: Redis connection lost and command aborted. It might have been processed.
Jul 10 11:19:23 refocus app/web.6: at RedisClient.flush_and_error (/app/node_modules/kue/node_modules/redis/index.js:357:23)
Jul 10 11:19:23 refocus app/web.6: at RedisClient.connection_gone (/app/node_modules/kue/node_modules/redis/index.js:659:14)
Jul 10 11:19:23 refocus app/web.6: at Socket. (/app/node_modules/kue/node_modules/redis/index.js:293:14)
Jul 10 11:19:23 refocus app/web.6: at Socket.g (events.js:286:16)
Jul 10 11:19:23 refocus app/web.6: at emitNone (events.js:91:20)
Jul 10 11:19:23 refocus app/web.6: at Socket.emit (events.js:185:7)
Jul 10 11:19:23 refocus app/web.6: at endReadableNT (_stream_readable.js:975:12)
Jul 10 11:19:23 refocus app/web.6: at _combinedTickCallback (internal/process/next_tick.js:74:11)
Jul 10 11:19:23 refocus app/web.6: at process._tickDomainCallback [as _tickCallback] (internal/process/next_tick.js:122:9)
Jul 10 11:19:23 refocus app/web.6: code: 'UNCERTAIN_STATE',
Jul 10 11:19:23 refocus app/web.6: command: 'ZREM',
Jul 10 11:19:23 refocus app/web.6: args: [Object],
Jul 10 11:19:23 refocus app/web.6: position: 0 }, .............
```

We have seen these 'UNCERTAIN_STATE' errors before as well. I went through other related issues and it seems a network problem, but since it looks like it is coming from kue, I just wanted to confirm whether we have some potential options to prevent this in future.

Start with the stack-trace locations in node_modules/kue/node_modules/redis/index.js, especially flush_and_error, connection_gone, and the EXEC/ZREM handling. Investigate how Kue handles Redis connection loss and whether the reported UNCERTAIN_STATE condition can be reproduced or documented. Done means establishing a specific preventive or recovery change, with tests or clearly documented operational guidance.

Written by the indexing model from the issue text.