Graphite + Statsd, different values on different time range

Question

While using StatsD and Graphite, I'm running into problems while viewing the same stats_counts.* metrics value in different time ranges:

As you can see from the graphs above, the same measured data is being viewed differently when picking a bigger time range.

I would've understand the loose of accuracy due to Whisper's storage-aggregation scheme utilised on older data, but this really doesn't explain how recent data is being displayed like it has a different value in different ranges of time.

Just for the record, my schema_aggregation.conf looks like this:

[munin]
pattern = ^munin\..*
xFilesFactor = 0
aggregationMethod = average

[min]
pattern = \.lower$
xFilesFactor = 0.1
aggregationMethod = min

[max]
pattern = \.upper$
xFilesFactor = 0.1
aggregationMethod = max

[sum]
pattern = \.sum$
xFilesFactor = 0
aggregationMethod = sum

[count]
pattern = \.count$
xFilesFactor = 0
aggregationMethod = sum

[count_legacy]
pattern = ^stats_counts.*
xFilesFactor = 0
aggregationMethod = sum

[logster]
pattern = ^logster\..*
xFilesFactor = 0
aggregationMethod = sum

[default_average]
pattern = .*
xFilesFactor = 0.3
aggregationMethod = average

and my storage-schemas.conf:

[carbon]
pattern = ^carbon\..*
retentions = 60:90d

[munin]
pattern = ^munin\..*
retentions = 10s:6h,1min:7d,10min:5y

[stats]
pattern = ^stats\..*
retentions = 10s:6h,1min:7d,10min:1y

[stats_counts]
pattern = ^stats_counts\..*
retentions = 10s:6h,1min:7d,10min:1y

[logster]
pattern = ^logster\..*
retentions = 60s:12h,10m:1y

# [default_1min_for_1day]
# pattern = .*
# retentions = 60s:1d

Any idea what might be wrong? maybe a configuration which I missed?

Answer 1

It looks like you're running into the regular Graphite behavior where it averages y-values when you have more x-axis data points than pixel in the graph (Since 2h of data has 720 data points). Does that also happen when you view the graph with &width=1000?