[metadata] metadata sidecar implementation

[davidnuescheler]

in situations where metadata is managed on a per page basis in a spreadsheet, depending on the size and volume of the metadata, it could be advantageous for pipeline execution performance to load metadata from a "sidecar"-file eg. .metadata.json from the contentbus instead of reading a potentially large aggregated metadata json.

[davidnuescheler]

in this implementation the last modified date could additionally depend on the last modifide date in sidecar file and therefore could be more fine grained which should have positive effects on caching and SEO crawling budgets

[hparra]

Would that include any page using canonical? Adobe.com is doing that for tens of thousands of pages per cloud.

[tripodsan]

Would that include any page using canonical? Adobe.com is doing that for tens of thousands of pages per cloud.

this issue is about an internal optimization. it will not affect how the metadata is rendered in the html page.

[hparra]

Ok. We have tens of thousands of lines in metadata and we will start adding custom json-ld for many them, some dozens of lines long. Hopefully we stay within limits and we don’t break anything.