Nagios Daemon Wishlist

From NagiosCommunity

Nagios Daemon Wishlist

If you have requests for new features or changes to the Nagios daemon, this is where you can submit them to make sure they don't fall off Ethan's TODO list for future versions. :-) It is recommended that you also propose changes on the nagios-devel mailing list, so other developers and contributors can comment on them.

Scheduled Downtime

• Recurring downtime
  • Suggested by Jeff Vier on 5/26/06 [1]
• Ability to remove downtime (not possible in version 2)
  • Suggested by Aki Tuomi on 10/29/07

Visibility Directive

• Ability to see hosts/services in the web front-end without getting notified
  • Suggested by Marlo Bell on 3/8/07 [2]

Alerting

• Summary alerts - when a host is down and there are other hosts downstream (unreachable), currently you can get a 'host unreachable' alert for each downstream host. I would like to be able to get a single summary alert that lists all the hosts (or at the very least the number of hosts) that are unavailable due to the downed host either as part of the original host down alert or as a seperate alert. Two messages are better than a flood of them if you have alerting for 'unreachable' hosts turned on or no alerts (other than the host down) if 'unreachable' is turned off.
• Audio Alerts - Currently Audio Alerts will be played on soft state changes, this should be changed to hard state changes or there should be a configuration option in the cgi.cfg to switch between hard/soft states.

SMP Support

• Nagios should support SMP hardware.

Database based Configuration

• I know that there's a good reason for how the configuration has evolved, but it would be really good (and would make scripting a decent configuration tool a lot easier) if the configuration was in a database table... perhaps someone could revive the Nagios-DB project, or maybe re-write the Nagios scripts to move away from the files?
  • Suggested by Jon Spriggs on 11th April, 2007
• Jon, what if there was some abstraction? I would propose a compile-time abstraction allowing some iteration of groups (ie all hosts) and specific values (ie second check for host XXX).
  • would allow outward extension for a DB config backend (skimming specific rows from a SQL backend), LDAP, or XML as the backend of the config.
  • I current can read hosts and services form a LDAP subtree with schema shared with DNS and DHCP, to some degree
    • by putting a "monitorGroup" attribute in the DNS entry, it should trigger at minimum a basic ping
    • by putting a "check-service" attribute in the DNS entry, I am adding Nagios checks to the DNS entry
    • by adding Nagios hostnames to a monitorGroup LDAP Object, I have redundant servers checking a group of hosts
    • schema is a bit bogus, but it passes a "nagios -v nagios.cfg", and I an interacting with Benoit (GoSA Project) for test-data
  • ChickenAndPorn on 1st January, 2007 (status update 2008-04-19).
• Documenting my progress on LDAP-Configured Nagios here starting 2008-04-20

Graceful Restart

• When Nagios configration are modified, we must be restart Nagios proccess, and Nagios will Re-scedule queen. Can Nagios to do something for NOT restart process when configration modified ?
  • Suggested by Sysit Chen on 26th April, 2007

Better network support

• Nagios has a rather simplistic idea of network topology. Instead of parent/child links, which are inherently one-way, it would be better to assume all links are two-way and say which hosts do routing. That would allow for networks with redundant connections. A simpler version of this is to allow for cycles in the parent/child graph.
  • Suggested by Peter Westlake on 30th April, 2007
• The possibility to make dependencies for a same interface. For example one interface with 2 services : one to check if the interface is up/down and one to check for example the traffic in/out on this interface. The traffic wouldn't be check if the interface is down but if we configure such dependencies in Nagios we have "circular dependency definition".
  • Suggested by Adeline on 05th May, 2008
• Only Really Simple Configurations have only one network interface per every host so that interface have only single IP address. It is not correct for the purposes of network monitoring system to say "IP" and mean "host".
  • More real configs often have more than one IP address per host. Even more, in our situation there are two management links to every host we monitor. Hosts are distributed geographically over a large area and we have to have two management networks to provide two different paths to each host. So every host have at least TWO IP addresses, and more, two paths.
  • Nagios does not allow to express in host dependencies simple situation when Layer2 topology is not a supergraph for Layer3 topology.
    • What if l2 topology is like this: NAGIOS<->SWITCH<->ROUTER<->... but l3 topology is NAGIOS<->ROUTER<->SWITCH ? Documentation does not say what topology should be used when building host dependencies. Perhaps, the idea of hostdependency is incorrect?

Automatic Network Discovery

• Nagios could have an auto-discovery feature to speed-up the configuration and maintenance of nagios-configs.
  • Suggested by Cero on 4th June, 2007

Group Memberships

• A group with no members should not cause a fatal error on Nagios startup. A warning would be fine.
  • Suggested by Lee Garner on 14 September 2007
  • I had this same itch a while ago. I just submitted a patch to the nagios-devel list. [3]
    • Moshe Hyzon on 23 October 2007

Keep Service in non-OK state until acknowledged

• There should be a service-option, which keeps the service in a non-OK state until the acknowledge button is pressed (good for logchecks etc.)
  • suggested by Martin Eichenberg on 08 Jan 2008

Instant recheck

• There should be a way to recheck a service "on the click" (not just requeue it) which should raise the check-counter as well and show the checkresult instantly to the operator.
  • suggested by Martin Eichenberg on 09 Jan 2008

Event Broker Module Error Messages

• NDOMOD was installed but someone accidentally mistyped the path to NDOMOD's config file. No error message appeared in nagios.log telling us anything like "NDOMOD config file not found".