Is single-write supposed to be slower compared to bulkSave?

[thearghyadutta]

In theory, it does seem that it is supposed to be slower since the API calls would add to the overhead time.

But in case of bulkSave, inserting a million documents in batches of 1000 takes 8 mins compared to batches of 1 or 2, which takes as long as an hour and 40 minutes based on a benchmark test I am carrying out.

Want to know if this time range of inserting a million documents seems normal to ArangoDB performance. My setup is a 3 node cluster.