Im running a simple python webserver on a raspberry pi (http.server) under the latest raspbian.
Now and again it fails to respond to clients though I cant see any pattern in the failures.
ps ax shows the program is still running.( ssh'ing to the pi)
Kill the program and restart it .. and its working fine and responding as it should.

Ive sprinkled the program with prints to a log file but I cant see anything meaningful in the log. It just seems to stop. The log file isnt large a few 100k. The pi's hard disk (yup :-) ) is 250G.

The server is only serving my local network, it gets only a few requests a day (say 6 ? ). It displays the temperature and humidity of the rooms in my house.(Gathered by other pi's and sent to the pi hosting the webserver and saved in text files) It has a couple of "buttons" to switch the boiler on and off.

Also running on the pi is Pi_Hole, an add removing DNS server which also serves up webpages using lighttpd and my own python program gathering and saving the humidity and temperature reports.

Can anyone point me in the direction of some bug detection techniques .. it may be that I'm using the module incorrectly of course .. So any suggestions of methods for getting more information would be great.

you need to insert print/log/debug statements into that python script. That is all I can suggest you at this moment. If you have no idea about the reason you need to put a lot of logs to find it.

Thanks pan64 .. but I've done that :-((

no, not yet. I mean if you could not find the reason you need to put more logs into the code. Probably on every function, probably even more logs (probably on every line) required to find out where is it stopped.
If you knew where you could also find the reason: why by also printing variables and other information related to the actual code.

To pan64 ..
Ive added logs to the start of all functions now, previously just on exit.
Now Ive just got to wait for it to fail :-))

Found the problem !
I ran the python script in idle3 and waited and waited for it to fail.
It turns out that the failure was due to a network failure in an other Pi and nothing to do with the webserver !
The script uses pigpio to access the state of a relay in the other Pi .. which is controlling the central heating boiler.
This controller Pi's networking has failed and hence the failure.

So ... on to that problem!

With regard to my original question adding lots of logs and running the script in idle3 helped solve the problem.
Most effective I think was running it in idle3 to get its error messages
Regards

that is something to learn: use descriptive error messages.
And another thing, if you wish you can implement some kind of timeout handling which can help you in similar cases.