Ad Widget

Collapse

zabbix_server: Debug log level can't provide useful data

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • sigil
    Junior Member
    • Feb 2014
    • 3

    #1

    zabbix_server: Debug log level can't provide useful data

    Hello!

    I've set DebugLevel=4 and it started to show many unusable info (from user POV). But sadly enough, it can't help me with figuring out why zabbix_sender have failed to send the data:

    26375:20140220:162425.089 Trapper got [{
    "request":"sender data",
    "data":[
    {
    "host":"xxxx",
    "key":"yyyy",
    "value":"100"}]}] len 129
    and then

    26375:20140220:162425.089 zbx_send_response() '{
    "response":"success",
    "info":"processed: 0; failed: 1; total: 1; seconds spent: 0.000110"}'
    The question is, why it spams with cyclic info about pollers but don't show failure causes? Where to look otherwise?
  • Wasabi-RSM
    Junior Member
    • Feb 2014
    • 11

    #2
    Originally posted by sigil
    Hello!

    I've set DebugLevel=4 and it started to show many unusable info (from user POV). But sadly enough, it can't help me with figuring out why zabbix_sender have failed to send the data:



    and then



    The question is, why it spams with cyclic info about pollers but don't show failure causes? Where to look otherwise?
    "Unusable Info" is a relative term - for your particular problem it might not help, but for a different error it might mean the world. Anyway, when you say the server failed to "send the data" - what exactly do you mean? The server doesn't necessarily "send" data, but rather "polls" the clients. So I guess the correct way of saying it is, the server asks the clients to send it the data.

    Comment

    • sigil
      Junior Member
      • Feb 2014
      • 3

      #3
      Originally posted by Wasabi-RSM
      "Unusable Info" is a relative term - for your particular problem it might not help, but for a different error it might mean the world.
      I've said it because messages like

      26381:20140220:162423.709 self-monitoring [processed data in 0.000086 sec, idle 1 sec]
      are hardly usable for anyone except the devs of server itself. And messages like

      26375:20140220:162425.089 zbx_send_response() '{
      "response":"success",
      "info":"processed: 0; failed: 1; total: 1; seconds spent: 0.000110"}'
      are usable for usual users. But since first ones flood over second ones without stop, it's not very good idea to mix them at one log level.

      Originally posted by Wasabi-RSM
      Anyway, when you say the server failed to "send the data" - what exactly do you mean?
      zabbix_sender have failed to send the data ...

      Comment

      • Wasabi-RSM
        Junior Member
        • Feb 2014
        • 11

        #4
        Okay, I understand - but what are you trying to accomplish? Are you trying to monitor a Windows host but no data is showing up, thus you resorted to enable debugging in the logs and now you cant find what you're looking for, or are you monitoring a windows host and only partial results are being returned?

        Forgive my ignorance, but your question (from my point of view) is like taking your car to the repair shop and saying "it doesn't work" with no further explanation?

        To quote a line out of one of my favorite movies: "Data, data, data - I cant make bricks without clay" - Sherlock Holmes.

        I'm by no means a Zabbix expert, but I'll try to help where and when I can - so maybe by giving more info other people might hit your post with a particular search and be able to support you better

        Comment

        • Wasabi-RSM
          Junior Member
          • Feb 2014
          • 11

          #5
          Sorry, one more suggestion, why dont you try setting your debug level to "3"? This way it will only show Critical, Error and Warning messages - if I understand correctly...

          Comment

          • sigil
            Junior Member
            • Feb 2014
            • 3

            #6
            As was said earlier, the problem is in that server logs "failed", but doesn't say about failure cause. What is not clear with that?

            Another not obscure fact, that I try to send some data and it gets into server. I don't ask you to help me figure out what is wrong with data or server. I ask to help me to make server say it about this for itself.

            Are the devs here?

            Comment

            • Wasabi-RSM
              Junior Member
              • Feb 2014
              • 11

              #7
              Okay mate, sorry about wasting your valuable time then - I hope you find what you are looking for.

              Originally posted by sigil
              As was said earlier, the problem is in that server logs "failed", but doesn't say about failure cause. What is not clear with that?

              Another not obscure fact, that I try to send some data and it gets into server. I don't ask you to help me figure out what is wrong with data or server. I ask to help me to make server say it about this for itself.

              Are the devs here?

              Comment

              • jan.garaj
                Senior Member
                Zabbix Certified Specialist
                • Jan 2010
                • 506

                #8
                Code:
                26375:20140220:162425.089 zbx_send_response() '{
                "response":"success",
                "info":"processed: 0; failed: 1; total: 1; seconds spent: 0.000110"}'
                Response is success, so don't worry :-)
                Check if host "xxxx" has item "yyyy", which must be enabled on your Zabbix server. Also check type of item (Zabbix Trapper; Zabbix Agent; ...), type of information (if you send string, but type of information is float => it'll fails, of course).

                You didn't gave a lot of details about your problem, but you are only complaining. Be calm and thankful if someone gave you some clues :-)

                BTW: you can see logged communications from Zabbix protocol in sever log file. Zabbix communication protocol should be lightweight, so it doesn't contains chatty information.
                Devops Monitoring Expert advice: Dockerize/automate/monitor all the things.
                My DevOps stack: Docker / Kubernetes / Mesos / ECS / Terraform / Elasticsearch / Zabbix / Grafana / Puppet / Ansible / Vagrant

                Comment

                Working...