why agent going to offline even when server is up and running fine
Team, We are facing agent disconnect issue frequently in one collector. we checked from network side and there is no packet drop loss and even in server, dynaTrace process is running fine. Just want to understand on which parameter collector consider agent as in disconnected state on linux server and what is pooling interval time for reconnect and also like to understand on what time interval agents sends data packet to collector. Agent - JVM. Thanks and Regards, Suraj Gupta
hi Suraj,
as Asad said, the Agents never intentionally disconnect from the Collector/Server.
if it does it's most of the times a network issue. you don't have a load balancer in use, right?
the logs you posted are from the Collector, but we would need the logs from the Agent, please. this log only tells us that the Agent is reconnecting for some reason. please use the System Information dashlet and go to down to the connected agents and collect the Agent logs from there.
this should give us some more hints about the problem.
thx, Christian
Suraj
The agent in the JVM never disconnects from the Collector. It is constantly connected to the collector. Can you check the agent log files to see any errors in it?
Thanks Asad, attached some logs, can you please have a look and let me know if you understand what exactly caused agents to be disconnected from the collector?
2015-08-17 14:48:17 INFO [AgentPeer] Hybris_HYBECMPRDAP09@hybecmprdap09:7775, agentId: 1d79f3d8, Version: 6.1.0.8154 uses:
2015-08-17 14:48:17 INFO [AgentPeer] System Profile : Production
2015-08-17 14:48:17 INFO [AgentPeer] Configuration : Default
2015-08-17 14:48:17 INFO [AgentPeer] Agent Group : GOE Hybris APAC
2015-08-17 14:48:17 INFO [AgentPeer] Agent Startup Time: Tue Aug 04 19:49:13 CEST 2015
2015-08-17 14:48:17 WARNING [ClassCacheReadingThread] Unable to resolve all required meta classes from class cache! Suppressing this warning for another 60000 ms.
2015-08-17 14:48:17 INFO [AgentListener] Using recovery information to reconstruct Hybris_HYBECMPRDAP09@hybecmprdap09:7775
2015-08-17 14:48:17 INFO [AgentPeerPool] added agent 'Hybris_HYBECMPRDAP09@hybecmprdap09:7775' to pool - number of agents: 16 Agent logs attached guys, can any1 say the exact issue here. i can see class cache issue here. if any1 understand this log then please let us know.
Thanks Chris,
I tried to collect log from agents but i am amazed to see its in unreadable format and logs are as well not updated properly.
Do you have any idea how to get rid of this issue? agent-logs.jpg Please have a look into the image which i attached Thanks, Suraj Gupta
wow, that looks weird! I'm wondering if there's maybe some very strange (asian?) character set in use?
do you maybe have access to the filesystem of those machines to take a look at those log files (<DT_install_dir>/log) if they look the same? you could also just collect an archive via right-click in the client, as it may be just an issue with the DT client.
anyway, I suggest to open a support ticket for this behavior.
best, Christian
Welcome the
Dynatrace Community Forums
Check out the Forum User Guide and Forum Guidelines to learn how to get started.Community Member of the Month
September 2017
Announcing Dynatrace's Community Member of the Month for September 2017, Tarun A.! Click here to read more! 
