Ids are 566, fetched documents are 566
--------------------------------Approach1-----------------------------------------
Normal Bucket:176ms
Ephemeral Bucket:158ms

Code:
            StopWatch stopWatch = new StopWatch();
            stopWatch.start("query");
            List<JsonDocument> foundDocs = rx.Observable
                    .from(ids)
                    .flatMap(id -> bucket.async().get(id))
                    .toList()
                    .toBlocking()
                    .single();

            stopWatch.stop();
---------------------------------Approach2-----------------------------------------
Normal Bucket: 382ms
Ephemeral Bucket: 386ms

Code:
            StopWatch stopWatch = new StopWatch();
            stopWatch.start("query");
            for (String id : ids) {
                JsonDocument jsonDocument = bucket.get(id);
            }
            stopWatch.stop();
----------------------------------Approach3------------------------------------------
Normal Bucket: 434ms
Ephemeral Bucket: 449ms

Code:
           StopWatch stopWatch = new StopWatch();
           stopWatch.start("query");
           Statement stmt = Select.select("*").fromCurrentBucket().useKeys(JsonArray.from(ids));
           N1qlQueryResult query = bucket.query(stmt);
           stopWatch.stop();
---------------------------------Approach4-------------------------------------------
Btw if I use spring repository, much more time is taken (3011ms).. I imagine its all the overhead of the json to entity mapping
Code:
           StopWatch stopWatch = new StopWatch();
           stopWatch.start("query");
           for (String id : ids) {
               TestEntity testEntity = testRepo.findById(id).orElse(null);
           }
           stopWatch.stop();