I have a simple ticketing system.
I need to show the number of tickets open for each client at the end of each week - since 1/1/2015.
This search will show the number of jobs opened for that week
...Project="MyClient1" | eval DateCreated=strptime(Created,"%d/%m/%Y %H:%M") | eval DateResolved=strptime(Resolved,"%d/%m/%Y %H:%M") | eval JobState=if(DateResolved>DateCreated,"Calculated Closed","CalculatedAsOpen") | where JobState="CalculatedAsOpen" | timechart span=1week count(JobState) by JobState
but does not count jobs open from previous weeks.
The query would need to evaluate DateResolved against the current timechart time to determine if it was still open.
A similar question is posted here https://answers.splunk.com/answers/78275/timechart-accumulation-of-all-events-from-previous-times.ht...
Without seeing your dataset, I can recommend that you would change your query a bit..
... | bin span=1w _time | stats count(JobState) by Customer
This assumes that you have a value for Customer declared, and that each event has this Customer field in the data.
My example search is probably misleading.
The JobState needs to be evaluated against the "point in time the TimeChart/Bin is currently in" to determine if the Job was open at that point in time.
Eg If DateResolved > CurrentDateInTimeChart then JobState=Open
This same problem is presented (and better articulated) here: