Number of impala queries running - graph

Question

Hi, I can find this kind of message in the Impala daemon logs:

impalad.pxlbig04.nz.thenational.com.impala.log.INFO.20190117-104024.29754:I0117 10:42:17.900411 9206 admission-controller.cc:515] Stats: agg_num_running=7, agg_num_queued=0, agg_mem_reserved=525.00 GB, local_host(local_mem_admitted=0, num_admitted_running=0, num_queued=0, backend_mem_reserved=75.00 GB)

Which shows agg_num_running=7

Why is this same kind of metric not available in the CM graphs ?

I want to know how close to the max running queries of 25 that I set in the default queue on Impala Admission Control.

BikramjeetVig · Accepted Answer

CM collects all the metrics from impala which are also available through the impala debug page.

Please take a look at the metrics mentioned here which you can use in CM to create graphs. Specifically "Impala Daemon Resource Pool Metrics" which you can use to create the graph for your use case that displays the number of running queries as follows:

Graph that displays the aggregate num of running queries for a pool across the cluster:

select total_impala_admission_controller_local_num_admitted_running_across_impala_daemon_pools WHERE poolName="root.<your_pool_name>"

Graph that displays the num of running queries submitted to a particular host for a pool:

select impala_admission_controller_local_num_admitted_running WHERE poolName="root.<your_pool_name>" and hostname="<host_address>"

You can further tweak these queries and explore the chart builder in CM to get the exact graph you want.

Sign up

Please use your Cloudera account.

Login to the community

Please use your Cloudera account.