Hello,
Duno if this has been already seen somewhere else but i d like to know if there is a way to avoid this problem :
Sometimes duno when exactely (thouhg i suppose it is when a server is crashing or if the agent is being stopped violently or something like that) we get on graph very high values, for example you can monitor a ntework link and have an average bandwidth used of 20 ti 50 Mb/s and suddenly you get a vry high value of several Gb/s.
The graph then becomes unusable as this big value way over the real capacity of the link just show a big peak and then the real graph (very low bandwidth compared to the peak) is then now almost invisible ...
Any idea how to avoid the server to not store very high values (out of range) recieved from agent ?
Thank you.
Duno if this has been already seen somewhere else but i d like to know if there is a way to avoid this problem :
Sometimes duno when exactely (thouhg i suppose it is when a server is crashing or if the agent is being stopped violently or something like that) we get on graph very high values, for example you can monitor a ntework link and have an average bandwidth used of 20 ti 50 Mb/s and suddenly you get a vry high value of several Gb/s.
The graph then becomes unusable as this big value way over the real capacity of the link just show a big peak and then the real graph (very low bandwidth compared to the peak) is then now almost invisible ...
Any idea how to avoid the server to not store very high values (out of range) recieved from agent ?
Thank you.
Comment