1 |
On 17/09/2013 07:42, Grant wrote: |
2 |
>>> Has anyone tried Nimsoft Monitoring? It's included at Soft Layer |
3 |
>>> which must mean a free license. |
4 |
>> |
5 |
>> No. IBM has a general strategy to "suck you in" |
6 |
>> so caveat emptor...... You really wan to install |
7 |
>> IBM binaries on any machine? (Think NSA). |
8 |
> |
9 |
> Nevermind! |
10 |
> |
11 |
>>> It looks like a substitute for Nagios. |
12 |
>> |
13 |
>> Nagios has been under numerous stresses for |
14 |
>> quite some time, for a variety of reasons, imho. |
15 |
>> Forking, Borking, and Porking out is what I see |
16 |
>> of Nagios; ymmv. |
17 |
> |
18 |
> I didn't realize that. |
19 |
> |
20 |
>> jffnms is well written, modular and quite responsive |
21 |
>> to the individual's (organization's) needs, imho. |
22 |
>> All in source code form. |
23 |
>> |
24 |
>> Last time I checked, there was a new (recent) ebuild |
25 |
>> for jffnms. Patches are easy to apply and I think |
26 |
>> (Gentoo) folks are starting to use jffnms much more. |
27 |
>> |
28 |
>> Check it out, most are happy with it, and find it |
29 |
>> easy (particulary with SNMP 1,2.3) to install and extend. |
30 |
> |
31 |
> It looks great, thank you for the recommendation. Have you used |
32 |
> munin? If so, do you think jffnms is a substitute or compliment to |
33 |
> that package? |
34 |
|
35 |
|
36 |
Munin and jffnms bear no real relation to each other. Yes they are |
37 |
similar in that both can draw graphs but that's about where the |
38 |
similarity ends. |
39 |
|
40 |
Munin's job is to periodically poll a device using whatever means is |
41 |
available and gather data from the device. The data is always in the |
42 |
form of a number - it measures something. The data can be anything you |
43 |
can generate a number for - logged in users, traffic through an |
44 |
interface, load, number of database queries. The list is endless. Point |
45 |
being, the device/computer/hosts reports it's own numbers to munin, and |
46 |
munin draws graphs. Munin does not record state, it has no idea what the |
47 |
state of something is. |
48 |
|
49 |
Nagios is a problem child, it does not do what people assume it does (I |
50 |
have constant fights about this at work). Nagios is a state monitoring |
51 |
and reporting engine (simply because this is what it does well and |
52 |
everything else it does it does poorly). Nagios will track if things are |
53 |
up or down, if you acknowledged the condition and when, who to notify |
54 |
when state changes (sms, mail, dashboard etc etc). |
55 |
|
56 |
What Nagios does poorly (despite this being it's advertised purpose) is |
57 |
getting state events into the system. It really really sucks at this and |
58 |
is coded from an extremely narrow point of view. Which explains the |
59 |
numerous forks around (they all implement vital real world features that |
60 |
Ethan refuses to commit). |
61 |
|
62 |
jffnms is something I don't use myself, but it looks like the same class |
63 |
of app as Nagios. Don't be fooled into choosing between munin and |
64 |
nagios/jffnms - they are not the same thing, not even close. Use both. |
65 |
|
66 |
|
67 |
-- |
68 |
Alan McKinnon |
69 |
alan.mckinnon@×××××.com |