[RedisLabs Users] Lots of Inactive Jobs

Hello - we have Kue running on an ElasticBeanstalk on AWS and connected to a 500MB Redis instance on Rediscloud - webservers are connected from Heroku. It’s set up with Redis Version Compliance 3.0.3.

Here’s how I have the connection set up.

jobs = kue.createQueue({
    redis:
        port: 18147
        host: "XXX.us-east-1-4.1.ec2.garantiadata.com"
        auth: "MyPassword"
    })

When I first deployed it, kue was processing the jobs perfectly. However, since deploying this we are seeing lots of jobs not even get processed and go right to inactive. They haven’t even been attempted to be processed. I looked at the other issues about inactive jobs but none seem to have applied. Does anyone know why these jobs are going direct to inactive and not failed/delayed? Let me know if I can provide any more information. Thanks

[peterporfy]

@jurgen7777 I don’t synchronize between processes but a different process will have a different connection and so the blpop issue won’t affect it. It is only to ensure that one connection is doing one blpop. Pretty much the same what you did in your PR. I tested it with 3-4 node processes with the test above. Btw @ghost is me too here I just merged my accounts. 😃

[tomandersen]

We had this issue with a kue on Heroku running about 10,000 events per day with RedisCloud we had delayed and other events going to inactive for up to 30 mins. It was fixed for us by adding watchStuckJobs() right after we created the Queue. It would be nice to see this bug fixed!