Update LKE API docs to reflect actual behavior.

LKE does not impose ASCII-only restrictions on the label name. The only
restrictions on the label are length and JSON validity.

Author: sibucan

openapi.yaml

@@ -15933,14 +15933,19 @@ components:
 
             Labels have the following constraints:
 
-              * Must start with an alpha character
-              * Must only consist of alphanumeric characters and dashes (`-`)
-              * Must not contain two dashes in a row
-          minLength: 3
+              * UTF-8 characters will be returned by the API using escape
+                sequences of their Unicode code points. As an example, the
+                Japanese character 'か' is 3 bytes in UTF-8 (0xE382AB). Its
+                Unicode code point is 2 bytes (0x30AB). APIv4 supports this
+                character and the API will return it as the escape sequence
+                using six 1 byte characters which represent 2 bytes of Unicode
+                code point ("\u30ab").
+              * 4 byte UTF-8 characters are not supported.
+              * If the label is entirely composed of UTF-8 characters, the API
+                response will return the code points using up to 193 1 byte
+                characters.
+          minLength: 1
           maxLength: 32
-          # Kubernetes does NOT allow underscores in resource names and we are
-          # passing this name to Kubernetes. Thus the following pattern:
-          pattern: '[a-zA-Z0-9-]{3, 32}'
           example: lkecluster12345
         region:
           type: string