Thank you @niketnilay for your response. I do find it odd to reverse the testing as you indicate and I see how it will work but it does not in my situation. The problem is not in how I have this coded but in how the code itself works. It is in that aspect that I think the failure is occurring. It seems that the condition match code looks at the Base search and not at the results of a Post-Process search. There may be something else going on as well as I changed one of my dashboard panels to use an embedded search and it stills shows the row when the end result of the search has a single value of zero.
Here is my embedded search:
tag=database tag=query action=failure | `find_NULL_fields` | stats count by _time, event_time, event_id, severity, src_ip, os_user, src_app, dest_ip, dest_host, db_name, db_schema, db_user, sql_command, object_type, error_code, action, query | bin _time span=24h | stats count by _time, db_user | where count > 14 | stats count as Violations
The idea is to collect the core events, bundle them into 24hr buckets and count the # events, then keep only those totals where the total is 15 or more then count how many of those are found. If I narrow my search to a specific time frame I will get 11 "core" events which does not provide sufficient total to provide a count over 14 so the end result for my single-value panel is zero. But there are 11 events returned so the $job.resultsCount$ parameter is not zero and the row displays.
I tried using the $result.count$ parameter which is supposed to allow me to use a value from the results set but that does not work at all and the $row1_show$ token is not generated. I get the same results if I use $result.Violations$ or $result.'Violations'$. This further deepens my suspicion that the condition match does not quite work as I am expecting.
Thank you again for your suggestion/ideas, although it doesn't solve the problem it does provide insight in looking at the problem in a different way. I appreciate that.
... View more