[jira] [Closed] (INFRA-12069) Mirroring JIRAs

["Gavin (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/INFRA-12069?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Gavin closed INFRA-12069.
-------------------------
    Resolution: Fixed

I see no harm , please carry on. As you say though, be careful with your scripting or you may end up hitting some limit that gets your IP banned. If that does happen file a new issue citing this one as reference.

> Mirroring JIRAs
> ---------------
>
>                 Key: INFRA-12069
>                 URL: https://issues.apache.org/jira/browse/INFRA-12069
>             Project: Infrastructure
>          Issue Type: Improvement
>          Components: JIRA
>            Reporter: Ewen Cheslack-Postava
>            Assignee: Gavin
>            Priority: Minor
>
> Our company does a bunch of work in the Kafka project and we're trying to figure out a better way to deal with project management across both Apache, our own OSS projects, and internal projects. We'd like to mirror JIRAs from Apache's instance into our own internal JIRA instance for our own tracking purposes.
> There don't seem to be great out of the box solutions for mirroring JIRAs, so I've written up a basic prototype script that does it. But I wanted to check with Infra before deploying anything to make sure this is kosher, and if so, also get some guidance on what frequency of syncing would be acceptable. We obviously don't want to impact Apache's infrastructure in any negative way.
> The script is pretty straightforward, and for reference, Kafka currently has < 5k JIRAs filed. It currently just hits the JIRA API a number of times to pull the basic JIRA metadata for each bug (in batches of 100 because that's the pagination Apache's JIRA Is currently limited to) and then does the rest of the processing against our own JIRA instance. I am using a very simplistic approach now, but we could easily make it more incremental to reduce the number of API calls (after initial mirroring, only search for new/modified JIRAs to apply updates). We also only care about mirroring the JIRA with basic metadata, we're not trying to copy over full comments, all changes, etc, just the title, description, status, and so on. So syncing anywhere from once an hour to once a day would be the maximum frequency we'd need.
> Is this ok to run against Apache's JIRA instance?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)