Collecting histogram diagnostics

Steps to enable collecting histogram diagnostics using the DataStax Enterprise Performance Service.

DSE provides histogram tables for this data:


Histogram	Details table	Summary table	Keyspace details	Keyspace summary	Global details	Global summary
cell_count	Y	Y	N	N	N	N
partition_size	Y	Y	N	N	N	N
range_latency	Y	Y	Y	N	Y	N
read_latency	Y	Y	Y	N	Y	N
sstables_per_read	Y	Y	Y	N	N	N
write_latency	Y	Y	Y	N	N	N

Note: These tables show similar information to the data obtained by the nodetool tablehistograms utility. The major difference is that the nodetool histograms output is recent data, while the diagnostic tables contain lifetime data. The data in the diagnostic histogram tables is cumulative since the DSE server was started. In contrast, the nodetool tablehistograms shows the values for the past fifteen minutes.

dse.yaml

The location of the dse.yaml file depends on the type of installation:

Package installations	/etc/dse/dse.yaml
Tarball installations	`installation_location`/resources/dse/conf/dse.yaml

Procedure

To enable the collection of table histogram data using the DataStax Enterprise Performance Service:

Edit the dse.yaml file.

In the dse.yaml file, set the enabled option for histogram_data_options to true.

# Column Family Histogram data tables options
histogram_data_options:
  enabled: true
  refresh_rate_ms: 10000
  retention_count: 3

(Optional) To control how often the statistics are refreshed, increase or decrease the refresh_rate_ms option in dse.yaml.

The refresh_rate_ms specifies the length of the sampling period, that is, the frequency with which this data is updated.
Optional: To control the number of complete histograms kept in the tables at any one time, change the retention_count parameter.