Marvel 2.0 Uneven measurement times result in odd graphs

monitoring

(Drew Town) #1

Starting today I started noticing my Marvel acting funny. The measurement times it takes is very uneven. With ever few minutes all of the readings stop and it shows as zero on the graph. If I zoom in to less than 4 hours the whole graph actually renders as a solid 0 line.

Any ideas on what I can do to fix this?


Marvel shows sporadic indexing activity - is this correct?
(Drew Town) #2

When looking at the node statistics this only seems to be happening on one of my nodes which is what is causing the weird indexing statistics.

Edit: this ended up resolving after I restarted the node that was causing the odd data. Definitely not an ideal solution as most of our nodes are under pretty heavy load. I will report back if this happens again.


(Chris Cowan) #3

The rates are calculated using a derivative pipeline aggregation. When the data drops or has significant gaps, then its impossible to do the calculation. This problem is amplified when your resolution increases because the gaps become more pronounced.


(system) #4