Emergency

i did pmm 1.6.0 upgrade. i think graph refresh has bug

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • i did pmm 1.6.0 upgrade. i think graph refresh has bug

    hello

    I did pmm upgrade to 1.6.0.

    I think graph refresh has bug at All dashboard.

    ex) mysqloverview

    i had setted config.

    Last 30 miuntes & auto refresh every 5s & interval 5s

    graph is jagged at auto refresh.

    It did not happen before 1.6.0.

    see attachment. thank you.
    Attached Files

  • #2
    Hi next1009
    If anything I believe the jagged lines in the graph will be due to the interval you set. For example you're asking Grafana to retrieve from Prometheus the last 30 minutes of history at a 5 second resolution level. The tighter the resolution then the more data points and thus the more jagged your graph will look. If you set the interval to something higher, for example 1m or 5m does it smooth the line? Here is my 1s view:
    Click image for larger version

Name:	1s-resolution.PNG
Views:	1
Size:	90.2 KB
ID:	50725

    Compare with the 1m interval:
    Click image for larger version

Name:	1m-resolution.PNG
Views:	1
Size:	57.0 KB
ID:	50726

    Comment


    • #3
      Hi
      but It did not happen before 1.6.0 on same config.

      jagged means : graph top point and time changed at auto refresh.

      see attachment. thank you.

      sorry. My english skill is not good enough.
      Attached Files

      Comment


      • #4
        hi

        I found diffrent scrape_time, at PMM 1.6.0 and PMM 1.5.3.

        PMM 1.6.0 is 30seconds and pmm 1.5.3 is 4 seconds. - see attachments.

        i think it's changed.

        So it seems strange when the graph is auto-refreshed (every 5s)

        am i right?
        Attached Files

        Comment


        • #5
          Originally posted by next1009 View Post
          Hi
          but It did not happen before 1.6.0 on same config.

          jagged means : graph top point and time changed at auto refresh.

          see attachment. thank you.
          Hi there - I'm not sure we are seeing your attachments? If it helps you can email my forum username @percona.com and I can upload them.

          Comment


          • #6
            hi
            i upload image again.
            can you see it?
            sorry for late.

            Comment


            • #7
              Hi next1009 , sorry for the delay in response. I checked in our 1.7.0 release on your scrape frequency question and it appears that we are scraping
              innodb_thread_concurrency
              in our Low Resolution scrape interval, which is every 60 seconds by default.

              Code:
              [michael.coburn@pmmdemo ~]$ for s in lr mr hr ; do echo "Resolution: $s" ;curl -sk https://10.91.136.32:42002/metrics-${s} | grep thread_concurrency; done
              Resolution: lr
              # HELP mysql_global_variables_innodb_thread_concurrency Generic gauge metric from SHOW GLOBAL VARIABLES.
              # TYPE mysql_global_variables_innodb_thread_concurrency gauge
              mysql_global_variables_innodb_thread_concurrency 0
              Resolution: mr
              Resolution: hr
              [michael.coburn@pmmdemo ~]$

              Comment


              • #8
                Hi
                but It did not happen before 1.6.0 on same config.

                jagged means : graph top point and time changed at auto refresh.

                Comment

                Working...
                X