[AMBARI-12718] AMS is doing a (Select *) for certain point in time metrics from Dashboard page

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 2.1.0
• Fix Version/s: 2.1.1
• Labels: ambari-metrics
• Assignee: Dmytro Sen
• Reporter: Dmytro Sen
• Created: 11/Aug/15 13:19
• Updated: 20/May/16 17:40
• Resolved: 11/Aug/15 16:09

Description

API calls for certain point in time metrics are fetching data for 1 day.

API:

http://hostname:6188/ws/v1/timeline/metrics?metricNames=dfs.FSNamesystem.MissingReplOneBlocks%2Cdfs.FSNamesystem.TransactionsSinceLastCheckpoint%2Cdfs.FSNamesystem.MillisSinceLastLoadedEdits%2Cdfs.FSNamesystem.SnapshottableDi rectories%2Cdfs.FSNamesystem.LastCheckpointTime%2Cdfs.FSNamesystem.TotalFiles%2Cmaster.Master.QueueCallTime_median%2Cdfs.FSNamesystem.ExpiredHeartbeats%2Cdfs.FSNamesystem.PostponedMisreplicatedBlocks%2Cdfs.FSNamesystem.LastWrittenTransactionId%2Cdfs.FSNamesystem.Snapshots%2Cjvm.JvmMetrics.MemHeapCommittedM%2Cdfs.FSNamesystem.TransactionsSinceLastLog Roll%2Cmaster.Server.averageLoad%2Cjvm.JvmMetrics.MemHeapUsedM%2Cdfs.FSNamesystem.PendingDataNodeMessageCount%2Cmaster.AssignmentManger.ritCount%2Cdfs.FSNamesystem.StaleDataNodes&hostname=namenode_hostnamel&appId=NAMENODE

Returned result on a cluster > 1 day old takes very long time.
The call should return 1 latest value for each metric vs all values.

Issue Links

• links to: Review Board
• links to: Review Board 2