{"p":"","h":{"iv":"ROXSYW+cfvEbFHu5","at":"ocxplSQjdRC3tXEtB/9/wg=="}}

Long-running background jobs (>60 seconds) are like potent denial-of-service attacks (DoS) that you cause yourself.

One or two is fine, but enqueue a few dozen and suddenly 100% of your background job capacity is consumed by just one class!

Break jobs into smaller parts.

2
Share
Share on Mastodon
Share on Twitter
Share on Facebook
Share on Linkedin
0
1y
Jamie McCarthy

@nateberkopec And/or, if using resque-pool or similar, assign a small number of workers to a queue dedicated to long-running jobs, leaving the majority of workers free for other queues

0
1y
Replies