Dynatrace plugin snmpv2 : JVM monitoring
Hi,
We are trying to use Dynatrace to monitor JVMs.
We want to see two different cases:
- Jvm down (crashed)
We uses the default alert “Application Process Unavailable (unexpected)” to monitor the down JVM. It works pretty good and we have the information we need.
- Frozen JVM (hanging)
We opened a case to monitor frozen JVM… we decided to use the default alert “Timed Out PurePath(s)” to monitor that.
We opened another case because of the plugin snmp by default which not supply the involved agents in the trap. The support tells us to use the community plugin, what we did.
However we have the same problem: the field “involved agent” isn’t available in the trap.
Have you any idear to add the filed “involved agent” in the snmp trap?
You'll find in attachment pictures of:
- Time out purepath alert
- Incident Rules
- Configuration of the plugin:
Regards,
Laurent
Hi Laurent
I will forward this to one of our engineers that worked on that plugin and hope he has some thoughts on this
Andi
Hi Laurent,
There is a known deficiency in the dynaTrace plugin interface which prevents plugins from getting complete information about the incident from the dynaTrace engine. Please see discussion about OOTB incidents here.
Let's check first if the “Timed Out PurePath(s)” incident provides incident's source. Could you please do the following:
- Set logging level of the SNMPv2 plugin to FINER, trigger "Timed Out PurePath(s)” and send me log file;
- Deploy the Extended EMail Action plugin and add to your actions list.
- Cut and paste content of the BodyHeaderExample.html file into the BodyHeader parameter of the plugin (see details here). The BodyHeaderExample.html file is in the src directory of the plugin jar file.
- Set log level of the Extended EMail Action plugin to FINER.
- Trigger the incident and send me the Extended EMail Action plugin's log file alone with the plugin's notification e-mail. Please also send me OOTB Notification e-mail which you have already configured in the actions.
Please feel free to contact me directly via e-mail.
Eugene.
Hi Eugene,
Thanks for reply.
As you can see in attachment, I already triggered the Incident Mail by default which is able to describe involved agents.
You'll find too a trap from the plugin v2 in normal log level.
I have set log level to FINER and I plan to install the Extended Email Action plugin on monday.
I'll let you know when this will be done.
I haven't your mail, so you can take mine by Andreas.
Laurent.
Hi Laurent,
Thank you for attaching requested log files and notification e-mails. Log file from the Extended Email Action plugin shows that the "Timed Out PurePath(s)” type of incidents has no violations (see screenshot here) and as of today there is no way of getting source of the incident. More specific, an object which is passed by the dynaTrace engine to the plugin's setup and execute methods and implements ActionEnvironment interface has no references to the source or sources of the incident. This is known deficiency (see discussion here) which has corresponding enhancement request filed by Andi in summer of this year.
Eugene.
Hello,
I am working with Laurent oin that topic.
Is it possible to "force" violations in order to gather agent name ?
Thanks a lot in advance for you reply.
Alain (GENERALI CSM)
