Portal showing stale asset data

[donaldgray]

Issue

Portal can be slow to show updated assets. If ingesting an asset the Portal can be slow to show changes, resulting in the image being available via orchestrator but a Portal banner is shown:

This image is current being ingested! If this is taking longer than usual, try the re-ingest process

Width and Height are also slow to update.

Causes

Is this caching at Portal or API level? At what level (memory or http)?

If the latter, can we use the cache invalidation hooks introduced in #803 to resolve? This might not work because we need to know when Engine is complete - is there any way to do this? Alternatively, if the caching is in-memory should the ttl be reduced right down as it's an inexpensive request.