uiDivStats Option to filter domains reported on by uiDivStats?

[eightiescalling]

I may have missed it in amtm but is there a way to ask uiDivStats to filter certain information when showing in the UI?

There are a couple of bits on the network here that mean any top 20 reporting is 70% occupied by 2 devices - one of which is queries to a single host on the network and the other is a multitude of domains all work related because of what seems to be a stupidly chatty laptop as it talks to various things at work.

An option to configure a filter to ignore *.blah.com in the UI reports would be handy.

 

[Jack Yaz]

I may have missed it in amtm but is there a way to ask uiDivStats to filter certain information when showing in the UI?

There are a couple of bits on the network here that mean any top 20 reporting is 70% occupied by 2 devices - one of which is queries to a single host on the network and the other is a multitude of domains all work related because of what seems to be a stupidly chatty laptop as it talks to various things at work.

An option to configure a filter to ignore *.blah.com in the UI reports would be handy.

I like this idea. Would you only want them excluded from the "top" charts, or all stats including key stats summaries, and the time graph?

 

[eightiescalling]

I like this idea. Would you only want them excluded from the "top" charts, or all stats including key stats summaries, and the time graph?

Others may have different views but I'd probably go with the "top" charts - as the quick stop to identify trouble makers and seeing differences of what normal looks like. The time graph of everything already goes beyond the "top" charts anyway and the underlying Diversion reports would remain unchanged.

How easy it is to make this all configurable would be another matter - I'm guessing setting the filter list via amtm and the UI then having options of apply or not is probably the right sort of balance.

Happy to help with testing against dev branches.

 

[Jack Yaz]

Others may have different views but I'd probably go with the "top" charts - as the quick stop to identify trouble makers and seeing differences of what normal looks like. The time graph of everything already goes beyond the "top" charts anyway and the underlying Diversion reports would remain unchanged.

How easy it is to make this all configurable would be another matter - I'm guessing setting the filter list via amtm and the UI then having options of apply or not is probably the right sort of balance.

Happy to help with testing against dev branches.

So far my idea is just using a text file with one domain per line, and I can feed that into the various SQL queries that generate the charts. Pattern matching could be supported by adding a wildcard, e.g. *.google.com would exclude every domain ending in google.com but not google.com itself, or typing google.com would only exclude google.com

Opted to go for a text file as I can easily expose that in the CLI menu, users can edit themselves via SCP if they prefer, and adding a textarea to the WebUI is easy enough as well. If there's a good uptake of the feature I can iterate and maybe consider making a table approach and guided CLI entry, but for a first release I think (hope) a raw text file would be sufficient?

 

[eightiescalling]

So far my idea is just using a text file with one domain per line, and I can feed that into the various SQL queries that generate the charts. Pattern matching could be supported by adding a wildcard, e.g. *.google.com would exclude every domain ending in google.com but not google.com itself, or typing google.com would only exclude google.com

Opted to go for a text file as I can easily expose that in the CLI menu, users can edit themselves via SCP if they prefer, and adding a textarea to the WebUI is easy enough as well. If there's a good uptake of the feature I can iterate and maybe consider making a table approach and guided CLI entry, but for a first release I think (hope) a raw text file would be sufficient?

Raw text file with the pattern matching behaviour you describe works for me