Closed dpb587 closed 10 years ago
Seems like a good idea to me
I realized generated IDs are allowed to use underscores as the first character which makes it difficult to avoid internal calls in proxy rules (not that GET
internal calls should be dangerous, but least access has been our convention so far). Given this was only a debugging change and there are other ways to easily lookup a document (e.g. kibana search for _id:YTLKANrZQ...
, /{index}/{type}/_search?q=_id:YTLKAnrZQ
), it's just easier to avoid hacking up the nginx proxy rules. Closing...
Probably the correct call at this stage.
For reference: http://stackoverflow.com/a/19898676/13238
Auto-generated ids are random base64-encoded UUIDs. The base64 algorithm is used in URL-safe mode hence - and _ characters might be present in ids.
Primarily for debugging, but it'd be convenient to allow direct access to documents by ID. The regex should probably just allow all
GET
s to/{index}/{type}/[^_].+
.For example... this document is being parsed incorrectly...