OpenNMS 101 - Module 9: The Page Sequence Monitor Video Lecture | A Complete Guide: OpenNMS - IT & Software

hi folks and welcome to another episode
of open any math 101 in this lesson
we're going to build upon the work we
did in the last module talking about
service assurance now one of the key
parts of open NMS is the ability to
monitor services service checks whatever
you want to call it a very very key part
of any network management system and
open and this is no different in that
episode we did a fairly simple monitor
based upon SNMP in this lesson we're
going to do one based upon a more
complicated monitor called the page
sequence monitor now I'm going to go
over this but I want to show you that we
do actually have documentation so if we
actually go to doc start of an MS org
and since I'm running 1901 I'm going to
click on this 1901 section for releases
go under the admin guide and over here
in this section you'll see in section 4
we have service assurance there's all
the different service monitors that we
do out of the box if you're used to
Nagios we've got an in our P based
monitor most of our monitors can take
the place of a lot of the IES check
scripts and other checks that most other
products do now we have a very simple
monitor we call the ICMP monitor that
does Bane you know is the remote system
providing the ICMP service we also have
a TCP monitor today I want to talk about
something called the paid sequence
monitor this was invented for customer
of ours who came to us and they wanted a
way to accurately monitor a fairly
complex web-based application and so we
came up with this thing called the page
sequence monitor now what the page
sequence monitor does is it does the
same kind of check that the HCP monitor
will do but it also adds the ability to
walk a series of pages so it's called
the page sequence monitor we follow a
sequence of pages each page can be
either a get or put so you can get a
page you can do a put to say log in to a
form and you can work your way through a
particular
web application and so that's what we're
going to do in this lesson so we have
this range of monitors and we're going
to use the paid sequence monitor and so
for exercise we're going to look at the
open and immense login procedure so
we're going to do is okay the service
that says can you log in to open a knife
and honestly do use this service we use
it to monitor demo open in my store
where we run in demo every so often
something happens prob about once a
quarter
it gets unresponsive and it's slow it
does actually come up but it may take 20
seconds or so the little page and so
what we do is that we have a criteria
where we can actually log in navigate a
few pages and log out and if that works
and everything works timely we assume
the application is running and so we're
going to create this service and it's
what we called open and MS - login we're
going to add it to the local host node
the one we've been using throughout all
these lessons and then we want to see
what happens so let me jump over here to
my open MS web console so again this is
1901 everything looks nice and green so
I'm going to go over to configure open
in a mess and manage provisioning
requisitions now there's a couple of
ways I can do this my rule of thumb is
if this service if whatever I'm trying
to monitor only exists on one or a
handful of nodes I don't go through the
effort of making a detector I mean it's
pretty easy to make a detector but I
don't make one because I don't want to
have to test that service on potentially
tens of thousands of devices when I know
it only exists on three so if I wanted
to I could just edit this class
requisition by clicking on the little
pencil icon edit the local host node and
down here under the interface section I
could edit that interface and I could
just add the service I could click add
service I could type in open an MS -
login save it rescan and that service
would be automatically associated with
the note but I want to show you how it a
way to do this with a foreign source
with a detector so I'm going to edit our
default foreign source here
and I'm going to add a detector someone
hit this ad detector button right there
now the texture name is going to become
the service name and it's very important
because this is a name that's going to
be used throughout the system for this
particular service so I'm going to use
capital o P n capital n MS - capital L
log in capitals and lowercase characters
matter so I'm going to call itself an NS
log in again for service names you can
use pretty much whatever you want
don't use spaces there's a reason which
I'll cover in another lesson later one
of you spaces has to do with filtering
filters - like spaces they use them as
own delimiter so what we try to do is
use dashes use underscore characters use
camel case but don't use spaces and then
for the class I want to choose HTTP I
can just type that in here notice that I
could use a CPS if this was an ACP s
service I happen to have not configured
SSL for this login yet so I'm just going
to use the HPP detector but if you are
using HTTPS for your open and a match
then you may want to you may want to use
HTTPS detector now if I just stopped
right here the open MS login would
connect to port 80 and basically match
anywhere I have HBP which is not what I
want so I need to refine this h-2b
detector with some parameters so I'm
going to add a parameter and the first
parameter I'm going to choose is port
I'm going to use the port because open
NMS runs on 89 80 and now that in my
environment that would probably be
sufficient I don't have any other
applications that are running on port 80
980 but oh excuse me
so I don't have any other things running
on port 89 80s so I could pretty much be
safe here but there are other
applications I believe you I want to
sneeze again I could have another
application that was running on port 80
980 so I want to refine this a little
more I'm going to add another parameter
and I want to choose a URL and the URL
I'm going to choose is slash opening
so now what it's going to do is it's
going to connect to the IP address like
right up here I'm running this on my
local network 152 21.1 37 port 80 980
slash oppan NMS so that will actually
kind of belt-and-suspenders so we'll
connect to port 80 980 and also look for
the open ms URL so I'm going to go ahead
and save that I'll save my changes to
the point source and then I'm ready to
synchronize this class requisition ok
yes for this it'll go through the scan
so now what it should do is go and find
it so let's go to my info nodes continue
to so as you can see in the time it took
me to do this it's discovered the open
MS login service here and so as you can
see it's not it's not monitored okay we
haven't monitored anything here so we
need to fix that
so we've discovered it we've associated
the open NMS login service with this
particular interface and now what we
need to do is actually create a monitor
for it remember discovery is different
than monitoring because if I connect to
port 80 and issue and h2b get and I get
a 500 error message back that tells me
that there's a web server sitting there
but it's down from a detection
standpoint I don't care I've discovered
a web server but from a monitoring
solution that would not be a adequate
criteria I'd have to say that that
service is now down because 500 is not
acceptable we need like a 330 or 400 or
not a 500 a 200 um 303 something in the
valid range of responses okay so how do
we create a monitor and jump back to the
slides here and here is our XML again
monitors are created in the polar
configuration file and here is our entry
for
particularly service so service name
notice capital ope NMS login this is
important
oh sorry so this has got a match
whatever we discovered it if you use the
lowercase which is perfectly fine just
make sure that this matches interval is
30 seconds this is again in milliseconds
for 30,000 milliseconds the default
would be 5 minutes
most people with paid sequence monitors
actually use a longer period of time 10
or 15 minutes because you don't really
want to pound your web application too
much not that one extra users should do
so but in many cases it's just not as
necessary that to test it as frequently
in this case I don't have to sit around
for five minutes during this
demonstration during this lesson and
waits always set everything with 30
seconds user-defined equals two remember
we don't use that but I always try to
set it there and status on if you're
working on a fairly large page sequence
monitor and you don't want to enable it
by default you can set status equal to
off go to lunch come back play with it
etc again I strongly recommend that you
run all of your configuration on a test
server development server I mean this is
a VM running open it a mess so it should
be pretty easy anyway so like most
monitors we have these parameters we've
got retries value of 1 notice that this
is a retry meaning there will be two
attempts the initial try and the second
retry timeout and sometimes you have to
set these longer in this case this
should be fine I said it's 5 seconds but
you can set these longer the more page
the more page sequences you're doing
skip this repository this just stores
the latency data and I'll talk about
this in two more lessons we talk about
data collection and then this is what
differentiates the paid sequence monitor
from other devices it's just page
sequence key and here's our page
sequence so the first thing says ok /
open ms / index.jsp 489 80 look for pass
so I'm going to go log out of here and
so I'm going to go to going to make sure
so this says I need to go to the IP
address slice open in ms index.jsp so
I'm going to go up here or type in index
dot JSP so as you can see this actually
we direct to the login page but that
doesn't matter what the base sequence
monitored we'll follow the redirects and
so it's looking for a capital P password
so I've got it right there there it is
there so this page sequence would pass
notice this virtual host a lot of web
servers aren't happy if you come to them
with just an IP address as you know web
servers can manage a number of different
domains we have one web server in the
office here that manages little vanity
domain so couple my personal birthdays
and you know there it's one web server
but there's a ton of domains on them so
we have to pass something to it to tell
it what the host is in this case we use
the mode label it does work in this case
local host but of course you'd want your
node label to match whatever you were
trying to do here okay so now that I get
to here I need to I want to do a post
notice the second page path says post
again we're going to go to this J spring
Securities check for eighty nine eighty
and we want to post and want to put in
some values how do I know this well
unfortunately what I have to do is
actually have to look at the source so
I'm going to view the page source set it
up so human beings can read it and if I
go and look here is this form again if
you're not familiar with HTML this will
be a little bit difficult but you can
see that this J spring security check is
a form in which we pass some values we
pass the J username here and the J
password here I just happen to know by
looking at this what these two are
again when you're looking at your
application you have to know that as
well you may end up having to talk to
your your web designer or just look at
the look of this source so if I go to
J's spring and security check again this
came from right here this action J
screen security check I'm going to set J
underscore username to admin and gender
score password to ending now I'm
assuming even though early early in this
class I told you not to leave the admin
password admin I assume you've done it I
assume it's still admin because I
haven't changed it online and when I use
that as the check again you would want
to use some sort of login that wasn't
valuable in the case of the customer who
requested this they were a help provide
health care provider and they were using
this to look up patients so we had a
fake doctor with who only had access to
a few fake patients so that we weren't
actually using real HIPAA compliant data
we just want to use this so I wouldn't
use admin admin as my test in the case
of our demo server the user name and
password uses demo demo because that's
what people use to access it so anyway
we want to login with admin admin so let
me go back to my screen here close out
that type in admin admin and remember
that so I'm logged in so let's go back
to my page sequence monitor so I paid
sequence monitor is actually going to go
to once this post isn't made it's going
to go to the open MS index dot JSP page
and it's going to look for capital L log
capital o out and if you see you don't
see it in the page anywhere but if I
actually look at the source which is
what open NMS is doing is doing gets and
it's looking at the source I can find
capital o log out so there it is it's
right there
it happens to be in the drop-down box
under here so that's where the log out
is so if it finds log out it says oh I
made it to this page now what I want to
do I'm logged in so in this case it says
go to slash open MS
last event and /index open them this /
depent / index and I got a page so that
seemed to work and so what is it going
to look for it's going to look for
capital e event capital Q queries which
is right there so that passes so I know
I made it to this stage and so on and so
forth I can add a whole bunch of
different pages here I think for the
customer we did this for they had 30 or
40 different pages they used and then
finally you want to log out we do not
want to use resources if you don't have
to so if I go over here and I'll look at
admin log out you probably can't see
this on the screen but it's /j
underscore Springer underscore security
underscore log out is the URL so I'm
going to actually go to slash of a term
acetate spread security log out and I'm
going to look for this input J username
that happens if you remember from
looking at the source that happens to be
part of the of the the login form for
openvms so that's it we brought one two
three four five pages and if we can do
this and then completes within five
seconds I assume that my service is fine
now remember every time we had a service
especially with every time we had a new
service we have to set what monitor to
use what class to use so when I embed
this in our polar configuration file but
we're going to make sure we add this
monitor service line at the bottom says
to use the page sequence monitor notice
we use the HTTP detector but we're going
to use the page sequence monitor because
again I want to beat this into your head
detection and monitoring are different
okay so rather than type I have you type
all this stuff in you can actually get
it from our training page this may move
but at the moment right now it's
www.opm.gov/health a training and I'm
going to go into the here's all the
slides and videos for these lessons get
put up here they're not there yet
because I haven't finished them guys
waiting the process is that kind of
like I can't put them there until I got
them done so but the configuration file
is so an exercise 9 and what I'm going
to do is I'm going to right mouse click
on polar configuration and copy that
link and now when I go back to my
command line so I'm on the open MS
server and opt up Internet see I can do
it W get - capital n which will replace
in existing files and I'm going to paste
in that link so now add my polar
configuration files we look at it go
down to the bottom remember we created
our own package in our earlier we had
our little class package for class so
I'm going to add my open MS login right
here and there's exactly what we had put
in there and then one go down at the
bottom and make sure the open and MS
login piece is there okay and we'll save
that and then I'm going to restart open
in the mess so the system control you
start yeah I work I always forget its if
it's because system control and service
reversed it but I think that's right
yeah okay so we'll let that run if you
ever want to see how it's doing you can
do a tail - F on top of it in this logs
manager that log and it should show you
as everything gets stopped and restarted
okay while that's going on I want to
review what we've done so we created our
login service detector and this was
where we went to the default foreign
source we added our open MS login and we
use the ACP detector with extra
parameters so you can use this for
anything that runs on the web we just
had an issue with our sarmale gbox
phenology happens to run on port 5000 so
I could if I had a Synology service I
could put it up phonology service HPV
detector port 5000 and then I could add
URLs and other parameters to make that a
little bit more detailed again you want
to detect you use as detailed as you
have to use
and then stop so again if I just in this
case how to put port 89 88 there may be
other applications running on point
eighty nine eighty and so I added the
URL next we we scanned the the we
synchronized who could we scan as well
and lo and behold we got this open an
immense login service notice that it's
not monitored because we had not created
a natural monitor for it so we created
to monitor we again matched open ms log
in here and match the open and mess log
in under the name we set some basic we
tries in time outs and then the page
sequence which is the important part
so let's see how that resource board not
sure it worked see one of the things
I've never been happy super happy with I
had not been super happy with yeah it
did work so let me go ahead do what's
the command line open an MS in opening
miss stop every since system seven came
in I'm not sure if it's our script or
it's what they're doing but we start you
could do it assist control stop open an
MS system which will start open MS in my
case I'm just going to sit here and use
the the regular scripts to do this so
while that's going back so at this point
in time in closets are there any
questions but again so while I use up
some time I'm going to talk about this
again remember when you're doing
detectors you could have gone directly
to this mode and added just typed in
open MS login and assign this service to
those you don't need a detector one of
the things I use or one of the things we
do in the office is we have a drink
machine that we monitor or with open an
MS it outputs the court of inventory
state to to a web page and then we use
the HTTP we use
of an HTP collector which I'll talk
about in a future lesson we use ACP
collector to go out and collect that
data now I only know that the drinks
service is what I think we call it the
drink service only exists on that
particular device so I don't necessarily
want a detector so all I did was went
into the service I went into the thing
in hid and hidden and typed in drinks
and that became my service I will look
for the manager no luck while I'm doing
this again I'm sitting there thinking
maybe I should just edit this down but I
want to show you this is like real life
and when you're working with open in the
mess it's very real so I want people to
become familiar with the logs and so
what I'm doing now is I'm looking at the
manager not logs so all the logs live
and opt open in this logs and as each of
the demons come up you'll get a log now
the first one to come up that takes a
while is this event D so once event D is
loaded so right now event D is being
loaded once it is loaded in running what
will happen is you'll see a whole bunch
of other services zoom by because they
take they're very very easy to load and
then it'll hang on jetty server now
jetty is what we use for our for our web
server so it's a Java servlet container
and so we use jetty for this and what we
will do is it takes a little while for
jetty to come up and when it comes up we
will know that it's time to log back in
as you can see this is the log up here
from the last time I restarted it so
they say invoking status on jetty server
and then as you can see status check
complete and it should say startup
complete so we will wait for that again
not sure what that thing is along but
anyway again when you discover a service
it doesn't necessarily mean you're
monitoring it and not monitored isn't
necessarily bad you can have services
associated with open and masters
specifically services kind of focused on
data collection they aren't monitored
you just don't there's no need to
monitor or something you're just doing
data collection work so not monitored
isn't bad but in this case it's saying
not monitored
because we never created the monitor
which is why we had to edit that polar
configuration file now a lot of stuff in
open MS is coming about where you can do
a lot of this through the GUI and we're
starting to work for the rest of 2017 to
actually make things like this easier
but for the time being you have to
actually edit your services using XML I
know it can be a bit of a pain but it
can be easier than writing a particular
script so okay so you can see all this
other stuff has started we got acne
alarm the event B and now we're sitting
here on jetty server so jetty server
looks like it's done we've got startup
complete which means we should be able
to log in it was kind of long okay so
I'm going to go over here click on
horizon should let me log in again I'm
using my tourists log in because you
shouldn't be using admin but we're going
to use it for this particular exercise
okay so we have three alarms more click
on these right now open MS login outage
identified connection refused connection
refused connection refused see these
reason codes notice it's come back up so
we basically determined all it went down
and went back up anyone in the class
guess what happened we were starting up
and it happens that the polar let's go
up here polar D starts before the jetty
server so we were already polling this
service before it came up and that's why
you'll get an outage remember there's no
magic in opening the mess it is magical
but there's nothing magic and so the
reason is that that I want you to know
where all this comes from so this reason
code connection was used tells me that
hey it could connect to it because I was
running I noticed very soon after it
came back up and we got all of this log
in outages alarms has been restored now
let's break it so I'm going to go to
configure open NMS configure users
groups in our controls
I'm figure users and I'm going to modify
the admin user and I'm going to reset
the password so remember it's currently
admin I'm going to change it to not
admin I'm actually using not admin but
you can use whatever you want so now
I've saved it to not admin and now if I
didn't do anything else nothing what
happens because I have to hit finish you
remember from the earlier lesson have to
get finished now the admin user cannot
login with admin so for those of my
front page here within 30 seconds we
should see the outage now notice the
last time we had an outage there were
three of them because it was on 127 0.1
121 72 21.1 37 and the ipv6 address
because we used a detector it detected
all of the all of the alders via all of
the open MS logins service on all three
because you could use all free to reach
it from the open MS side and I just
edited the node and actually forced open
an MS that's login to be on a particular
interface it would only be on one again
whether you want them on all or none
I'll leave that a few but that kind of
explains why that happened now as you
can see we have three alarms again and
39 seconds so we've got outages the icon
I happen to put an address in on my node
so it's showing up on the map so this is
Pittsboro North Carolina where we are
and that's 220 Chatham business Drive
where I'm doing this and so that's why
that icon change color and so now we can
go go to the info nodes you will see
that the open MS login services down on
all three interfaces
now the reason code is different it says
failed to find the log out in page
content so since I can see it fells if I
log out I know that it got to this page
but it didn't get to this page so what
happened was it was able to get to index
dot JSP so the login so that
near clue hey maybe the login isn't
correct and it had it broken if it had
said failed to find information on the
event index then we would know that it
had broken there so this gives you some
kind of clue as to where your problem is
now if I click on the service you'll see
that the page sequence for this thing is
displayed in there so you can kind of
look you don't have to go digging into
the configuration file nor do your
operators your operators you know and
kind of figure out oh ok this is where
it happens we do odds you skate anything
with password and so as you can see the
admin password here has been changed to
X and X so very top secret now before my
downtown model kicks in I want to go and
set the admin user back to admin so I'll
configure users modify the admin user it
reset password and then and then
remember to hit finish and within 30
seconds we should see these outages
clear and that's kind of it so notice
that everything is kicking in we've got
nodes with outages here so we've got an
outage there we've got an overall
service availability because we don't
have the open and MF login service in
one of our categories it hits the
default category which says three of
those serves you down we've got notices
that there's some issues we've got an
icon that's changed color we've got
alarms so Bing Bing Bing everything's
all upset that you had an outage now
I'll refresh outages cleared the alarms
it cleared notices it cleared and the
node has turned back to green and so
that's it now I want to tell you
something else about so let's go and we
go back to the slides just for one
second so we did that restart it open to
match remember that didn't work for some
reason so you might want to stop and
restart remember that we got an outage
it may fail at least or depends on how
fast your server is if you have a lot of
devices in your database instead of just
the one and you are
very fast system Jenny might come up
before the services polled but if the
service happens to be polled before the
jetty servers come out you'll get an
outage again we changed the admin
password
we got an outage and the thing was
different so we got outages we've got
alarms we got notices perfect now what I
want to talk about one last thing is
notice that in doing this you had to
kind of know a lot about your web
application we had to dig in to the
source and look for forms and things
like that
we have a better solution called the
selenium Monitor so I don't know if many
people have known selenium if I clicked
on swimming here so selenium HQ org is
the website and it's a it's a browser
test suite and what we can do is you can
actually create using the selenium IDE
you can actually go through here test
your web application using selenium
output that tests open as a mess and
import it in such a way that open a mess
can then run the selenium test there's a
number of caveats considering versions
and things like that and it's also not
documented so if I go to the
administrator guide now look selenium
Monitor is not yet documented I am
planning on doing an entire lesson in
open MS 102 the next class I do force
the selenium Monitor but at the moment
right now if you want to play with it
kind of work your way through the
through the wiki and you can feel free
to ask in the discussion groups or on
chat epidemics calm in the open image
discussion anyway I hope you found this
useful
umm we're going to try in the next
lesson I'm going to try and toggle watch
stuff together but I just wanted to show
you this lesson that we do have monitors
they do get more complex but they're
very very useful anyway thanks again see
you next