Due to an object placement issue in the cluster some HEAD and GET requests may result as a 404 error status althought the objects are actually in the cluster. We are investigating
Update(s):
Date: 2018-05-18 08:01:33 UTC 0.62% remaining
Date: 2018-05-15 07:48:11 UTC 1.1% remaining
Date: 2018-05-14 07:12:22 UTC 1.2% remaining
Date: 2018-05-11 16:00:24 UTC 1.40% remaining
Date: 2018-05-11 13:59:53 UTC 1.45% remaining
Date: 2018-05-09 14:17:54 UTC 1.8% remaining
Date: 2018-04-27 08:55:16 UTC 3.8% remaining
Date: 2018-04-26 07:08:44 UTC 4.3% remaining
Date: 2018-04-23 15:40:52 UTC 5.3% objects are still misplaced in the cluster (note that all misplaced objects aren't unreachable as only one replica is needed to be correctly placed to get a positive response from the cluster)
Date: 2018-04-13 14:47:14 UTC a robot rebalanced the data more than needed, leading to misplaced objects. We stopped it and we are trying to optimize the objects placement recovery with the less impact possible on performance