TODO

* ALERTS based on input from multiple targets
  Wouter Prins <wouter * null0.nl>


* ASSIGN BLAME - HALFWAY PING
  - A method for pinging the host in the middle between here and host B
    (use traceroute for figuring out which one it should be)
    This graph combined with the graph for here-B could show in which half of the link the problem lies ... 
    Russell Stuart <russell-debian * stuart.id.au>

* TOP10
  - show a list of the top 10 (most whatever graphs)
    gmourani * privalodc.com
    < they are willing to pay for this feature - tobi >

* UPTIME
  - define update via snmp pointer per device
  - possibility to call an external script
    -> generic uptime plugin?

* ATTENTION
  allow to define a threshold rule by looking at
  two average medians and take action when threshold triped

* ALIASES
  allow to have atarget which points to a different target
  only targets with host are considered

* ALERTS
  make 'active alerts' (whatever that means, have to think
  this through) visible (eg. different colour) in the CGI menu
  - suggested by Cornel Badea <cornel *sbhost.ro>
  
  Targets/alertee: syntax to remove an address from the Alerts/to list
  (maybe 'alertee = -user@somewhere'? This breaks real addresses
  starting with '-'...require quoting for those?)
  - from James Bouressa, <http://lists.ee.ethz.ch/smokeping-users/msg01667.html>

* REMOTE EXECUTION
  generic remote probe
  - a possibility for basefork-derived probes to reuse the same
    SSH connection with shell for() loops for all the pings to a given
    target

* GENERIC EXEC PROBE
  - almost every probe has a different way of calling system(), exec()
    or similar. This should be in an inheritable module.
  - the module should also support extra commandline arguments 

* RRD 
  configurable RRD parameters per target?
  - suggested by Leos Bitto,
    <http://lists.ee.ethz.ch/smokeping-users/msg01632.html>