Problem/Motivation
When normalizing collection requests we should be able to leverage the cached data from the normalization of the individual entities.
Proposed resolution
If a resource (entity) has been already normalized in the past, then reuse it during the normalization process. Incrementally expand the cache entry for a resource (an entity), just like Render Cache already does (see #16).
See a more detailed explanation of the performance improvements in #108. This is a summary.
| Scenario | Resp (avrg) | Resp (max) | Resp (min) |
|---|---|---|---|
| No Cache | 459 ms | 817 ms | 436 ms |
| With Cache | 191 ms | 925 ms | 174 ms |
Remaining tasks
None.
User interface changes
N/A
API changes
None.
Data model changes
None.
Release notes snippet
JSON:API now caches resource (entity) normalizations, resulting in a typical 2⨉ speed-up. The improved performance is especially noticeable on responses with many resources (entities), typically when using lots of includes. Notably, using different combinations of sparse fieldsets does not cause different cache entries; all requests for the same resource uses the same cache entry, regardless of the sparse fieldset requested.