This patch divides the query timeline script into multiple modules
for better maintainability and to help writing unit tests.
The patch also improves parsing of utilization values from the
query profile, as the order and name of counters may change within
the profile, to ensure appropriate mappings between plotted readings,
initially the profile is traversed to find the correct mapping.
Change-Id: Id9d62a799d838876509686a75ecea778b2c72cc6
Reviewed-on: http://gerrit.cloudera.org:8080/20588
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Add functions in CatalogUsageMonitor to monitor and report the
catalog usage of the tables have the longest metadata loading
time(Including maximum, median, 75th-ile, 95th-ile, 99th-ile time).
Set default tables loading metrics capacity to 100.
However, there might be a problem here because we only keep the
capacity size to 100. For example, there might be case like a
table has higher median loading time but has lower Maximum
loading time which cannot make itself to the Top-100. For now,
we will ignore case like that because we are aiming to find
the tables with maximum longest loading time.
Add the sorted table in Catalog server web-ui. The loading
time is sorted by the maximum from load_duration metrics. But
users can sort by other metrics in catalogd debug UI.
Testing:
- Add end-to-end test for webpage to verify the label and text
exist in catalog debug page. Verify all fields are in JSON response
- Launch Impala and activate some tables to see the table loading
time shown successfully on the catalog debug UI page.
Change-Id: I9305a867d7053cde9acc42dae6e47ee440f1a8bf
Reviewed-on: http://gerrit.cloudera.org:8080/14600
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
This patch reuse the JS function introduced in IMPALA-6966,
which can convert Long size value into human readable size,
to render the "Peak mem. consumption" column in the Query
details's Backends page.
Change-Id: I04afb4091bb8b6bc9dedfeceaf9284a8c65b16a1
Reviewed-on: http://gerrit.cloudera.org:8080/10523
Reviewed-by: Philip Zeyliger <philip@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
