Announcement

Collapse
No announcement yet.

Suggestion: cpanel "addon" to send notification emails to customers when LVE limit exceeded.

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

  • #31
    CPU and IO do not produce "faults" when their limits are exceeded. They simply slow down the account.

    So there needs to be another way to allow customer to disable those "triggers" for the notification email, other than number of faults.

    Comment


    • #32
      whatever on screenshot will be there.
      We will not have separate thresholds for CPU/MEM/NPROC, just one field for all of them.
      And while right now CPU/IO don produce faults, we can still collect & count as faults any time user hits the limit, and increment count in such cases.

      Ability for user to disable notifications --> we are not ready to implement that. This is something that will be part of one of the next releases.

      Comment


      • #33
        All sounds great to me -- really looking forward to this. Thanks VERY MUCH Igor!!

        - Scott

        Comment


        • #34
          Agreed this will be very handy.

          Comment


          • #35
            +1
            Would also be very interested in this feature.

            Whether automated or manually run - I an considering adding Inode limits, but would like a report to see how these new limits would affect existing users first.

            Comment


            • #36
              Im using the new feature that is the beta lve manager, to send alerts to the administrator (not sending customer emails yet).

              I have it set to send notifications every 5 minutes (for now) and a requirement that there is at least ONE fault in order to send emails.

              Here is an example of an email that I am being sent:

              > Dear Administrator,
              > Following accounts experienced issues in the past 5 minutes
              >
              > ================================================== ==================
              > | Username | Domain | Name | CPU | IO | Memory | EP | NPROC |
              > ================================================== ==================
              >
              > | example | example.com | --- | 6 | 289 | 349.2 MB | 14 | 13 |

              What is confusing is that none of the numbers listed above are hitting any of our maximum limits:

              CPU usage (SPEED): 320 %
              Number of cores for LVE (nCPU): 8
              vMEM: 2048 MB
              pMEM: 1024 MB
              EP: 40
              nPROC: 100
              IO: 1024 KB/s

              These alerts are nearly useless if they dont tell us what fault the user tripped. And the data doesnt show that ANY fault was tripped. Can this be clarified or fixed?

              Thanks!

              - Scott

              Comment


              • #37
                Hello Scott!

                We will improve notifications in the future releases.
                Parameter that was exceeded by the user will be marked in notification e-mails, for example:

                Dear Administrator,
                Following accounts experienced issues in the past 25 minutes

                ================================================== ==================
                | Username | Domain | Name | CPU | IO | Memory | EP | NPROC |
                ================================================== ==================

                | example | example.com | --- | 6 | 289* | 349.2 MB | 14 | 13 |

                Also, I suggest you not to set notification period less than 10 minutes, because it can cause false notifications due to average calculations of values of load parameters.

                Comment


                • #38
                  Hi Anton. Has this been improved yet? When we receive these notifications, we still are not being told what resource the customer exceeded.

                  We did change the notifications to 30 minutes.

                  - Scott

                  Comment


                  • #39
                    1. Please fix that cron job because its sending every minute messages to clients and admin.
                    2. Please add into email body the hostname because they are specifying only the domain name (doh... on what server its that account ?)
                    3. Please make the email files "_default.txt" and "_custom.txt" do not be overwritten on the next update.
                    4. Thank you for your great job!

                    Comment


                    • #40
                      I simply cannot believe that over one year later, we still do not have notifications that are useful when a customer hits a fault.

                      I turned on notifications, and sure enough, CloudLinux continues to generate completely useless notifications, like this:

                      Code:
                      Dear Administrator,
                      
                      Following accounts experienced issues in the past 1 hours
                      
                      ============================================================================
                      
                      | Username |     Domain     | Name | CPU | IO | IOPS | Memory | EP | NPROC |
                      
                      ============================================================================
                      
                      | someuser | someuser.com | ---  |  0  | 1  |  0   | 4.9 MB | 0  |   0   |
                      
                      ============================================================================
                      
                      The values are average values for the account, for the period. Metrics with * are the one that were limited.
                      Seriously? What value does this email have? It hasnt told me anything useful, other than this customer had a fault with "something" during the past hour.

                      Improving fault notifications is way overdue. Any update on this?

                      - Scott

                      Comment

                      Working...
                      X