Worker automatically cleans job registries every hour

[selwin]

Version 0.5.0 introduced StartedJobRegistry and FinishedJobRegistry. Unmaintained, job registries could grow to really large sizes. This is reported by @cosminstefanxp in #519 .

This PR introduces a feature where workers will clean up job registries every hour without having to run out of band maintenance scripts, simplifying worker deployments.

If there's no complaints, I'll merge this in a few days.

Fixes #519.

[cosminstefanxp]

It looks really good. Thanks for this. I'd suggest though that you add a way of specifying the cleanup time. Something like a property of the Worker that can either be changed when extending the class or set at creation.

[selwin]

I agree, I also thought about adding a user configurable "maintenance_interval" argument that can be specified during Worker.init() . That can be added in a later pull request.

Sent from my phone

On May 23, 2015, at 7:14 PM, Cosmin Stefan-Dobrin notifications@github.com wrote:

It looks really good. Thanks for this. I'd suggest though that you add a way of specifying the cleanup time. Something like a property of the Worker that can either be changed when extending the class or set at creation.

—
Reply to this email directly or view it on GitHub.

[nvie]

Thanks, @selwin, this is very well done! I've renamed the maintenance_date to last_cleaned_at, to reflect the nature of the state value a bit more accurately (see ec0e047).