Runtime Views

We provide a set of d$ views that expose the runtime state of the system.

d$druidservers view

Show the Historical Servers available for querying.

col_name	data_type	comment
druidHost	string
druidServer	string
maxSize	bigint
serverType	string
tier	string
priority	int
numSegments	int
currSize	bigint

d$druidsegments view

Show the Druid Segments of each DataSource.

col_name	data_type	comment
druidHost	string
druidDataSource	string
interval	string
version	string
binaryVersion	string
size	bigint
identifier	string
shardType	string
partitionNum	int
partitions	int

d$druidserverassignments view

Show which Druid Segments of are served by which _Historical Server_(s)

col_name	data_type	comment
druidHost	string
druidServer	string
segIdentifier	string

d$druidqueries view

Shows details about the Druid Queries that were issued during Query Execution. Maintains a history of the last sparkline.queryhistory.maxsize queries. By default this value is 500.(Note this is not a Spark property, by a Java system property)

col_name	data_type	comment
stageId	int
partitionId	int
taskAttemptId	bigint
druidQueryServer	string
druidSegIntervals	array<struct_1:string,_2:string>
startTime	string
druidExecTime	bigint
queryExecTime	bigint
numRows	int
druidQuery	string

StageId, PartitionId, and TaskAttempId

use them to associate this Druid Query to a Spark Task-Stage-Job

druidQueryServer

is the Historical/Broker server that was sent this query.

druidSegIntervals

is the list of (Segment Id, Interval) that this Druid Query ran on.

startTime

Time when Druid Query was issued, in ISO DateTime string

druidExecTime

Time for Druid Server to respond with first byte.

queryExecTime

Time for all the result data to be received by the Spark Task.

numRows

num rows received from Druid.

druidQuery

The Query sent to the Druid Server.