@isoutamo Thanks for your reply. you have used actual data in make results ? Not sure if thats the best for my case as the data comes in dynamically to splunk for the DB Connect. I have managed to get the individual events as desired by using mv commands. | eval zipped=mvzip(mvzip(vm_name,event_name),event_time) | mvexpand zipped | eval data=split(zipped,",") |eval vm_name=mvindex(data,0) | eval action=if((mvindex(data,1) like "Logon"), "logon", "logout") | eval assign_time=if((mvindex(data,1) like "logon"), mvindex(data,2), "NA")|eval release_time=if((mvindex(data,1) like "logout"), mvindex(data,2), "NA") | streamstats window=1 current=f values(release_time) as machine_release_time by vm_name, user_name | streamstats window=1 current=f values(release_time) as machine_release_time by vm_name, user_name | where assign_time !="NA" OR machine_release_time!="NA" OR machine_release_time!=" " | where action="logon" | convert timeformat="%Y-%m-%d %H:%M:%S" mktime(assign_time) as assign_time_epoch | convert timeformat="%Y-%m-%d %H:%M:%S" mktime(machine_release_time) as machine_release_time_epoch |eval session_time=(machine_release_time_epoch-assign_time_epoch) | eval session_duration=tostring(session_time, "duration") However, now I m stuck to put this in a chart using bucket command. The span doesnt work as expected. my session_duration is of type string and it has values like 00.37.35, 02:45:30 and so on. I am trying to put the chart using buckets so I can see machines logged on for 1-5 mins, 30-1h, >2 hrs and so on. any google search for string to date and use in chart leads me to strptime and it doesnt work as expected. because it converts to epoch and the time is unreadable in chart visualization, any way I can show the graph with HH:MM:SS with buckets with different span like 5min , 30m 2h etc.. TIA
... View more