Several different things:

1. That online article is correct to specify turning off/un-checking Mixed Queue in Rulerunner's settings. If you have done it and restarted Rulerunner afterwards, you should (start to) see only one empty but running batch at any given time. In your screenshot, many of these empty/running batches are showing, but were created on various dates. So, you need to clean them up to avoid confusion.

2. However, Rulerunner will always create an empty/running batch to wait for the next (group of) file(s) to show up, which is its current design. If you want it to behave differently, you need to open an ER, to which I believe many have for this very same situation.

3. Furthermore, the very first (group of) file(s) today will be imported into the very last batch dated yesterday, as it works as explained, unless you use a scheduler to stop/(re-)start Rulerunner.

I will keep the default behavior of rulerunner creating an empty batches and if business decides to change it, I will open an ER.

With the current rulerunner design, can I use my version of NENU application to delete old dated running batches?

Rulerunner should not be leaving multiple tasks in the running state. This is not an enhancement. Something is wrong. The most common cause of this issue is that the skipsamebatch option is not set to 1 as explained in the Technote. However there are other causes. If you have rrs logging turned on in Rulerunner, check the log to see if there is an error or crash of some kind. Also recognize that changing a running batch back to pending (or any other state) involves the Rulerunner thread making the request back to the Taskmaster service, which writes it back to the database. If something is interfering with one of those handoffs, then that could also cause this issue. See if you can capture an event where the batch was not set back to pending and check the Rulerunner and tms logs to see if there might have been a problem getting the database record for the batch set to Pending.

After changing one of the batch from running to pending status, I saw I had two jobs in Pending status for a while. Around 10 mins later that jobs status was changed back to running automatically.

In my tms log file I have found few 'Execute Failed' and 'will close connection' error message. Please see attached screenshots.

I would not assume that the connection error is specifically the cause of this issue unless you can show direct time correlations between both sides of the logs. The expectation if you were getting intermittent connection errors would be that any task might give you those errors. The fact that it is specifically the Import task in the running state suggests that the issue is something else.

You state that the batch eventually went back to the running state but of itself this is correct. As the import task polls for files, it actually goes back and forth between pending and running. However, most of the time it will be running, so periodic refreshes of the Job Monitor will show it as always running.

If you have multiple running Import tasks, this would indicate that tasks are not completing in some way. Use your rrs and Rulerunner logs to see what happened.

I don't know the frequency for rulerunner to look for the file in the directory but every time it does, rulerunner writes into rrs log files so I have same messages repeatedly written into the file.

Other that that I am unable to find any errors or unusual things.

Is there any settings in rulerunner which I can change and tells rulerunner to keep the batch in pending mode until it has next file to process? right now, after certain time it changes pending batch into running even though it doesnot have any file to process and creates a new batch in pending status.

The batch checks to see if there are any files by running rules. To run rules it must be in a running state. So the answer to your question is no, a batch can not check for files but remain in the pending state.