Provide a setting to run a spider against a specific archive

[leewesleyv]

Ideally we want to be able to optionally invoke the spider to run against the most recent archive by passing a CLI option or environment variable (picked up in settings).

[wvengen]

Since the spider knows where to locate object storage, it is relatively easy to figure this out. If a system talking to Scrapy needs to figure this it by itself, it needs to know container storage details.

As an addition to this feature, one could also perhaps provide a date/timestamp to locate the last archive before that (but that may depend on the configured storage path, so could be tricky).