Hello,
This does indeed look like a performance problem, normally on the Enterprise Manager side.
Howver, it also depends - do you see the same gaps at the same time for a whole agent, or even all agents, or does it only affect one metric on one particular agent ?
If only one agent is affected, there could be a problem with that particular application - it could even be an issue with the JVM itself.
We would need to know what performance guidelines you have followed, how many metrics you are reporting to the Enterprise Manager.
Hopefully if you have followed performance guidelines, you will have raised the heap of the Enterprise Manager from the default 1GB max of heap to 2GB, possibly 4GB.
You have to remember that the Enterprise Manager needs to collect and store data at 15 second intervals, and if it is overloaded, it won't be able to collect or store the data in time, and that is when you will see the gaps as you present.
So the normal thing would be to reduce load either by taking agents away or even trying to stop some agents reporting too much data.
Also increasing heap can help, or if you only have a standalone Enterprise Manager, you might have to consider a cluster, provided you have the hardware available to host this.
We would want to see the Enterprise Manager logs (including perflog) to give any sort of informed recommendation.
Thanks,
David