Description
When the Logstash Versioned Plugin reference build fails, it stops building and publishing for all documentation for all products on elastic.co. This is a regression.
Regression
In the past, a failure in the VPR build created a branch and PR to point out the point of failure and make it easier to fix. (See #20.) This was a good thing because it gave us a chance to fix the problem without affecting building and publishing all Elastic documentation.
Why this causes problems
The VPR build runs every 6 hours as an automated job. Any failures are now passed up to the elastic-docs build which runs every 30 minutes. If the VPR build fails, every elastic-docs job fails until somebody addresses the problem, meaning that nobody can get their docs built and published in the meantime. These failures often happen outside of US work hours, and I wake up to a fire drill.
Fixing the VPR is not obvious or straightforward, and well-meaning people in other time zones often try incorrect fixes that I have to revert. I have written some internal documentation on what needs to happen to fix the VPR so that I'm not the only one who can do it. But it would be much better if we didn't bring down the entire doc build, right???
Proposal: Go back to solution implemented in #20.