[ https://issues.jboss.org/browse/HAWKULARQE-81?page=com.atlassian.jira.plu... ] viet nguyen edited comment on HAWKULARQE-81 at 4/6/17 1:36 PM: --------------------------------------------------------------- [^promgen-analytic-12hr.png] gives a snapshot of raw metrics tallies during a 12 hour run of 230 pods. Grafana dashboard url: http://hawkular-grafana-analytics.cloud2.jonqe.lab.eng.bos.redhat.com/das... (data may be pursed automatically after 14 days) Expected raw data points per 15 minute interval: 30 Observations: - As seen in the top graph raw metrics dip below 20 - there are metrics definitions but 0 data for some pods - I suspect the dips were caused by pod restarts but that is not the case. I randomly selected 2 pods, 1 with restart and 1 without restart and graph their data counts. Next step: - store metrics id by pod name as oppose to pod UID for easier cross references with output of oc commands - reduce the number of pod to see if there are still dips - I just notice a flaw in my Python analytic tool - it fetches *all* metric definitions (including metrics of pods that are from previous test runs) Need to only get metrics for pods from current run. was (Author: vietn): [^promgen-analytic-12hr.png] gives a snapshot of raw metrics tallies during a 12 hour run of 230 pods. Grafana dashboard url: http://hawkular-grafana-analytics.cloud2.jonqe.lab.eng.bos.redhat.com/das... (data may be pursed automatically after 14 days) Expected raw data points per 15 minute interval: 30 Observations: - As seen in the top graph raw metrics dip below 20 - there are metrics definitions but 0 data for some pods - I suspect the dips were caused by pod restarts but that is not the case. I randomly selected 2 pods, 1 with restart and 1 without restart and graph their data counts. Next step: - store metrics id by pod name as oppose to pod UID for easier cross references with output of oc commands - reduce the number of pod to see if there are still dips -- This message was sent by Atlassian JIRA (v7.2.3#72005)