Provide sampling interval and offset in the plugn_sets output

[nichamon]

There are no query commands that show the sampling interval of a set, so the patch makes plugn_sets to provide this information.

[tom95858]

Can we rename this command sampler_sets? It doesn't apply to all plugins only sampler plugins so the name is misleading. Are there test scripts that use this command? If so, we could still support it, but make it 'deprecated'

[baallan]

Since this is new,
a) why not just call it 'list_sets' since we don't care how sets get produced
b) why not make it provide the same output (and output options, i.e. json) as ldms_ls, instead of reinventing the wheel.
c) might be a good opportunity to merge some of the ldms_ls formatting code into the ldms(d) library as utility functions

[tom95858]

@baallan, the command is not new, @nichamon is just adding the sample rate to the output. I'm not saying that what you suggest might not be a good idea, but let's do it as a separate task when we have time.

…

On Thu, Aug 21, 2025 at 8:52 AM Benjamin Allan ***@***.***> wrote: *baallan* left a comment (ovis-hpc/ldms#1923) <#1923 (comment)> Since this is new, a) why not just call it 'list_sets' since we don't care how sets get produced b) why not make it provide the same output (and output options, i.e. json) as ldms_ls, instead of reinventing the wheel. c) might be a good opportunity to merge some of the ldms_ls formatting code into the ldms(d) library as utility functions — Reply to this email directly, view it on GitHub <#1923 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABVTPXCDNV4DCEXZLHRVJ6T3OXMJHAVCNFSM6AAAAACD5O7FXWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTEMJQHEZTKOJTGY> . You are receiving this because your review was requested.Message ID: ***@***.***>

-- Thomas Tucker, President, Open Grid Computing, Inc.

[baallan]

Two additional comments:

1. this is a breaking change if anyone has scripted around the output of 'echo plugn_sets |ldsmd_controller ...', so plugn_sets should take an option that controls the output format, with the default being the original behavior.

2. this addition is wholly redundant with ldms_ls -v, so why are we doing it at all? just use ldms_ls.

[tom95858]

@baallan, I'm not sure I'm getting your point because ldms_ls doesn't tell you which plugin/config is responsible for producing or the sample interval.

[tom95858]

@nichamon, @baallan to get back to my original point, plugn_sets does not tell you which plugin is responsible it tells you which configuration is responsible. If meminfo had two configurations, there would be separate entries for each configuration.

[baallan]

So do we track which plugins subscribe or publish also?

I suppose I can see the value of the original plugn_sets, and the value of dumping all the configs of a plugin instance (which is another command, right?), but I don't see the value added by overloading plugn_sets with just the interval data. Just use ldms_ls -v, which does already provide that.

[tom95858]

@nichamon, this @morrone's patch provides a place to access the information provided here. What do you think?

[morrone]

@nichamon, this @morrone's patch provides a place to access the information provided here. What do you think?

@tom95858 I am not aware of any patch that I made that would change how Nichamon is doing things. Could you please elaborate on your thinking?