Agent reconnects happens in milliseconds intervals

Description

This problem has been reported by a customer:

It started basically with agents which could not connect anymore:

This log looks fine as the interval increases for every try. But after some days, it looked like this:

Could it be that we somehow hit a limit after all the retries ?

Environment

None

Gliffy Diagrams

Activity

Show:

Technical UserMarch 13, 2017 at 1:24 PM

SUCCESS: Integrated in

inspectIT - Integration #259
INSPECTIT-2335: Agent reconnects happens in milliseconds intervals (ivan.senic: d4c974fd8f80628fb4cdbdfd0970e812b363b817)

  • (edit) inspectit.shared.all/src/main/java/rocks/inspectit/shared/all/cmr/service/IKeepAliveService.java

  • (edit) inspectit.agent.java/src/test/java/rocks/inspectit/agent/java/connection/impl/KeepAliveManagerTest.java

  • (edit) inspectit.agent.java/src/main/java/rocks/inspectit/agent/java/connection/impl/KeepAliveManager.java

Ivan SenicMarch 8, 2017 at 11:04 AM

Maybe it's not a problem then there.. Do we have any loop somewhere?

Marius OehlerMarch 8, 2017 at 10:43 AM

I'm resetting the noConnectionCount, now, but this should not be the problem of the reconnection in milliseconds interval. Basically, this is very strange because the reconnection is triggered by the coreServiceExecutorService in a 2.5s interval. So even this counter problem is existing and there is an attempt to reconnect in each iteration, the minimum duration between two attempts should be 2.5 seconds.

Marius OehlerMarch 6, 2017 at 10:09 AM

Yes, I'll have a look on this.

Ivan SenicMarch 6, 2017 at 8:20 AM

Seams like when we hit the maximum we are not resetting the noConnectionCount back to zero and then it occurs every time.. We should fix this asap.. Marius would you have a look?

Fixed

Details

Assignee

Reporter

Integrator

Sprint

Fix versions

Affects versions

Priority

Created March 3, 2017 at 9:31 AM
Updated March 13, 2017 at 1:38 PM
Resolved March 13, 2017 at 1:38 PM