[
https://issues.apache.org/jira/browse/TC-270?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jeremy Mitchell closed TC-270.
------------------------------
> GET api/cache_stats needs to query summarized influx data rather than raw data
> ------------------------------------------------------------------------------
>
> Key: TC-270
> URL: https://issues.apache.org/jira/browse/TC-270
> Project: Traffic Control
> Issue Type: Improvement
> Components: Traffic Ops, Traffic Stats
> Reporter: Jeremy Mitchell
> Assignee: Jeremy Mitchell
> Priority: Minor
> Fix For: 2.1.0
>
>
> The following API queries raw influx data causing it to be slow and thus it
> times out:
> GET api/1.2/cache_stats
> That API endpoint currently runs 2 queries like such:
> series query:
> SELECT sum(value)*1000/6 FROM "metricName" WHERE time >
> '2017-05-03T09:22:00-06:00' AND time < '2017-05-04T09:22:00-06:00' AND cdn =
> 'cdn1' GROUP BY time(60s), cdn ORDER BY asc
> summary query:
> SELECT mean(value), percentile(value, 5), percentile(value, 95),
> percentile(value, 98), min(value), max(value), sum(value), count(value) FROM
> "metricName" WHERE cdn = 'cdn1' AND time > '2017-05-03T09:22:00-06:00' AND
> time < '2017-05-04T09:22:00-06:00' GROUP BY time(60s), cdn
> Those queries should leverage the summarized data and look like this instead:
> series query:
> SELECT sum(value)/count(value) FROM "monthly"."metricName.cdn.1min" WHERE cdn
> = 'cdn1' AND time > '2017-05-03T09:22:00-06:00' AND time <
> '2017-05-04T09:22:00-06:00' GROUP BY time(60s), cdn ORDER BY asc
> summary query:
> SELECT mean(value), percentile(value, 5), percentile(value, 95),
> percentile(value, 98), min(value), max(value), sum(value), count(value) FROM
> "monthly"."metricName.cdn.1min" WHERE cdn = 'cdn1' AND time >
> '2017-05-03T09:22:00-06:00' AND time < '2017-05-04T09:22:00-06:00' GROUP BY
> time(60s), cdn
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
