-
Notifications
You must be signed in to change notification settings - Fork 15
Compactions
This dashboard contain various statistics relating to compaction.
You can choose the following templates for the dashboard:
- Cluster - the Cassandra cluster name
- DC - the Data Center
- Instance - the node name
- GroupBy - datacenter/instance/cluster (not relevant for all 'per Node' graphs)
- Keyspace - the name of the keyspace (applicable only for keyspace pending compaction)
Describes statistics about compaction processes per node. The compaction graphs describe both the compaction tasks that reside in the thread pool as well as the actual sstable compactions that have taken place. Please note that there is not a 1:1 relationship between the compaction tasks and the actual compactions taking place on disk.
Please also note that changing the "group by" parameter above will have no affect on the graphs and they will remain per node. The same goes for Keyspace filter which will not filter compactions for the specific keyspace sstables.
Graphs:
- Total Compactions Completed / 5 min - the total number of physical sstable compactions which were completed during periods of 5 minutes
- Completed TP Compactions Tasks / 5 min - the number of compaction tasks within the thread pool that have completed during periods of 5 minutes
- Bytes Compacted / 5 min - the total number of bytes compacted during periods of 5 minutes.
- Pending TP Compaction Tasks - the estimated number of pending compaction tasks to perform.
Describes the pending compactions per different keyspaces. For each keyspace you choose in the template above, a separate graph will be created, describing only those sstables compactions related to this keyspace.
You can choose to group by DC, cluster and so on, and this choice will sum up the total number of currently pending compactions across the DC, cluster and so on.
Graphs:
- Keyspace Pending Compactions / 5m - the number of the current pending compactions per specific keyspace, sampled every 5 minutes