In which way did the failure manifest? From your mapping it looks like the field might be called netflow.flow_duration. The stepSeconds should be smaller than the size of the buckets that will be used in the date histogram aggregation on the x-axis of the visualization. If the value is greater, gaps might appear in the chart.
@weltenwort, this is my error:
Caused by: java.lang.ClassCastException: Cannot cast java.lang.Double to java.lang.Long
at java.lang.Class.cast(Class.java:3369) ~[?:1.8.0_131]
at sun.invoke.util.ValueConversions.primitiveConversion(ValueConversions.java:247) ~[?:?]
at sun.invoke.util.ValueConversions.unboxLong(ValueConversions.java:121) ~[?:?]
at org.elasticsearch.painless.PainlessScript$Script.execute(def stepSeconds = doc['netflow.flow_duration'].value;
Have you tried using double stepSeconds = doc...?
@weltenwort, netflow.flow_duration is a float.
@weltenwort, Does the duration_steps field has to include start and end times in an array? I should be able to calculate that field before it comes into elasticsearch.
That depends on whether you want the call to be considered "ongoing" in those start and end buckets or not. I would guess that you want to include them.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.