The same raw metric data looks different in Ganglia Plots because the underlying RRDTool applies Normalization and an Average Consolidation function on the data. This acts like a lowpass filter and remove spikes. Overall this emphasizes the underlying average in plots and can make noisy data more interpretable. Those are cases where in ClusterCockpit the plot is completely colored, when multiple noisy lines are overlayed.
Solution: Offer either a mode where the RRDTool algorithm can be applied to the metricstore, or change in general how the metricstore processes data.
The same raw metric data looks different in Ganglia Plots because the underlying RRDTool applies Normalization and an Average Consolidation function on the data. This acts like a lowpass filter and remove spikes. Overall this emphasizes the underlying average in plots and can make noisy data more interpretable. Those are cases where in ClusterCockpit the plot is completely colored, when multiple noisy lines are overlayed.
Solution: Offer either a mode where the RRDTool algorithm can be applied to the metricstore, or change in general how the metricstore processes data.