Clients signal strength history

RouterOS Scripting
1

Hello everyone!

I am looking for a script that can be imported to a program or webpage where you can see every client in the networks signal strength history, so if a client disconnects for a weird reason you can look at the history charts and see that it’s bad or good signal!? Is this something you can do in say the dude? I’ve seen something like this but in a commercial program that is not freeware, would be fun to make something that everyone can use :smiley:

Regards,
Mattias

2

Not sure if this will be of help but check the link below as I was looking for a similar solution and if it does solve your request please post back an update so we can all benefit?

http://forum.mikrotik.com/t/snmp-graphs/23058/14

3

in The Dude, you can execute any RouterOS command. so if you can obtain some info from Treminal - you can draw it in The Dude

4

Unless I have Dude 3.6 incorrectly configured, it shows average signal for the last 5mins and I cannot click and zoom into any graphs,
I find dude OK for current time monitoring and had to setup individual charts for monitoring customers who had signal issues but also with these I am unable to click and zoom into any graphs.
ClientSignal4.png

5

is your signal called ‘pppoe-out1’? (O_o)

6

I was using pppoe but have now changed to wlan, this has resulted in more information being displayed, AP+Freq,Signal,TX,RX,data rates but still cannot click and zoom into graphs,

7

We are using this:

  1. each client MT periodically runs a script which sends RX/TX levels (and some other infos too) to the log. Each MT box is configured to send some logs (script info) to remote syslog server
  2. a syslog server (linux box with syslog-ng) runs a script which parses messages from MT boxes and stores the interesting values into database. To spare disc space older values are merged into day averages so we have one months of detailed data and several years of day averages for each MT box.
  3. via HTML interface we can see RX/TX signal history for each client.
  4. we use nagios (resp icinga) to check if there is enough clients active (i.e their last message is not too old) on each AP. This way we can detect APs with bad antennas, wifi cards automatically. We run tests which analyze database data and detect signal drops (total average signal for all clients, etc) and bad trends too.

Regards
D. Toman

8

v4 of the Dude has more detailed graphs

9

each client MT periodically runs a script which sends RX/TX levels (and some other infos too) to the log. Each MT box is configured to send some logs (script info) to remote syslog server

Can you post the script here,

thanks

10

The script is nothing special. It is something like this one:

#USAGE:

- SCRIPT

/system script add name=sigmon policy=read,write

- LOGGING: remote action should be set to send syslog to x.x.x.x:514

/system logging add action=remote topics=script,info prefix=mt

/system logging add action=remote topics=script,debug prefix=mt

- SCHEDULER:

/system scheduler add name=sigmon on-event=sigmon interval=20m

/interface wireless monitor wlan1 once do={
:log debug ("sig:Rx/Tx:$"signal-strength"$"tx-signal-strength")
}

Comments:

  • if you have large amount of clients, don't use NTP on them - it is better if the boxes don't send messages to syslog server in the same time.
  • this way the signal will be sent to syslog and not stored to MT box log (if you want to log it locally you can use :log info instead of :log debug
  • yo have to have a script/app on remote syslog server which will parse the messages and store the TX/RX signals in the way you prefer.


    Regards
    D. Toman
11
  • if you have large amount of clients, don’t use NTP on them - it is better if the boxes don’t send messages to syslog server in the same time.

Thanks - but just thinking if the clients are not time synced, when investigating any issues then it makes it more difficult,

this way the signal will be sent to syslog and not stored to MT box log (if you want to log it locally you can use :log info instead of :log debug

I would at present prefer to log locally as i find Dude 4 ok for general network monitoring,


Thanks “normis” dude v4 has a much improved graph zoom and helps a lot

12

Thank you everyone for your answers! I will look up this Cacti program and look some at the new Dude v4.. Will post any progress I make!

13

Hi,
How about You progress, already done?
i setup CatiEZ and now try find solution to get log from MT router. any information link would appreciated

mo

14

any Syslog server, and then setup remote logging on MT