Note: In the following message, I use the term "Network" as to what
intermapper calls a network: a yellowish bubble hanging off from a
device. Not an actual IP network.
Personally, the way I would envision multiple probes on a device would be
like treating that device as a switch, with multiple networks hanging off
of it. Each "Network" would actually be another probe. Labels on these
"networks" should be customizable via the return values of the individual
probes (dynamic labels).
The other way to do it would be to write a custom probe that would in tern
call all the others (at the moment, a probe can't call another probe).
I don't think either of these methods would change the fundamental way that
intermapper works, and would just be minor tweaks to the code. Obviously
the first one (networks having probes) may require a bit of code, since I
don't know how exactly "networks" are defined.
Dynamic labels are something I've been screaming for ever since I first
tried to use Intermapper (about 3 or 4 years ago). As well as the ability
to update variables used in labels from custom probes of any
flavor. Hopefully, when the database version comes out, I can force custom
probes to finally be able to be useful by just whacking at labels in the
database. Then multiple probes per device will be a non issue, as I can
just set the values to what they need to be. Would be nice to just glance
at a label and see the problem rather than having to open up the status to
see that a printer is out of paper, or toner... or maybe it'll be handy to
see how many clients are connected to an AP without having to open the status.
- - - Jon
At 10:19 AM 5/8/2008, you wrote:
Richard E. Brown wrote:
> --- Michael Luzadder wrote:
> How close are we to getting a single icon to represent multiple probes on
> the same device? (like a mail server; one probe for host resources, another
> probe for snmp, another for http, etc..)
> --- end of quote ---
>
> I'm actually working on a specification for this now.
>
> My current thinking is that a single icon in a map would represent a
> "collection" of probes. The collection would test a single IP address.
Any of
> InterMapper's probes could be added to a collection (including other
> collections). These probes would be executed sequentially, once every poll
> interval.
My problem with this proposal is that it requires a significant change of
code. The existing architecture is one vertex
--> one device --> one probe.
Rich is proposing one vertex --> one device --> multiple probes.
In my opinion, one vertex --> multiple devices is easier to implement
because there is no change in polling behavior.
The notion of multiple "subdevices" is easier to shoehorn into the
existing schema for IMDatabase than devices with
multiple probes.
In IPv6 land, devices have multiple IP adddresses as a normal state of
affairs. If you wanted to test a MAIL server over
both IPv6 and IPv4, you would still need two icons in Rich's proposal. In
my proposal, the devices could be grouped
under a single icon even though they have different addresses. In Rich's
proposal, you would be required to poll the
mail server, host resources, http server *all* at the same poll interval.
In my proposal, you could set different poll
intervals for the different "subdevices" of the one visual icon.
Just my thoughts.
Bill Fisher
Dartware, LLC
____________________________________________________________________
List archives:
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [EMAIL PROTECTED]
____________________________________________________________________
List archives:
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [EMAIL PROTECTED]
