Current Value Not Resetting when Device goes down

[kelbyc]

I have created thresholds which are working correctly, however, when the device goes completely down Cacti is not setting the data source value to 0 or to blank. Instead it is retaining the last good value until the device becomes available again. I'm trying to use THold to alert me when these devices go down but since the current value never resets the low threshold is never crossed. The dead host notifications will work as well but I only want to monitor certain devices on my network and not all of them. I have over 300 devices and at times there is a flood of emails if I turn dead notifications on. Is there a way to make the current value change to 0 or blank if the host is dead or is there a way to only monitor certain devices and send dead host notifications to those rather than all devices in Cacti?

KC

[noname]

Probably Cacti doesn't regard unknown data as 0 (like as MRTG's "unknaszero" option).

... is there a way to only monitor certain devices and send dead host notifications to those rather than all devices in Cacti?

Though I didn't verify them,
1) manage plugin
2) nmid + nmidUpDownMonitor
See also: http://forums.cacti.net/viewtopic.php?p=215701#p215701

- or -

3) Try this modification to thold plugin (for thold-0.4.4-41 or 43)

Code: Select all

> diff -u thold/includes/polling.php.org thold/includes/polling.php
--- thold/includes/polling.php.org      Mon Feb 14 23:04:46 2011
+++ thold/includes/polling.php  Wed Aug 31 20:12:03 2011
@@ -174,6 +174,11 @@
                                        }
                                }
                                $host = db_fetch_row('SELECT * FROM host WHERE id = ' . $id);
+                               if (api_plugin_is_enabled('monitor')) {
+                                       if ($host['monitor'] != 'on') {
+                                               continue;
+                                       }
+                               }
                                if ($host['status'] == HOST_UP) {
                                        $snmp_system = '';
                                        $snmp_hostname = '';
@@ -288,6 +293,11 @@
                                        continue;
                                }
                        }
+                       if (api_plugin_is_enabled('monitor')) {
+                               if ($host['monitor'] != 'on') {
+                                       continue;
+                               }
+                       }

                        $downtime = time() - strtotime($host['status_rec_date']);
                        $downtime_days = floor($downtime/86400);
@@ -351,6 +361,12 @@
                                if (plugin_maint_check_cacti_host ($host['id'])) {
                                        continue;
                                }
+                       }
+                       if (api_plugin_is_enabled('monitor')) {
+                               $monitor = db_fetch_cell('SELECT monitor FROM host WHERE id = ' . $host['id']);
+                               if ($monitor != 'on') {
+                                       continue;
+                               }
                        }
                        $failed[] = $host['id'];
                }

If you enabled monitor plugin, alerting will be skipped for the host whose "Monitor Host" checkbox is off.

[kelbyc]

Thanks for the help! The polling.php solution worked great. Now I am only monitoring the devices I need to monitor instead of all of them. My only last request would be how I can set it up to only alert if the device has failed for 2 polling cycles instead of just 1. I tried to attach the modified polling file but I couldn't seem to find an extension that the BB would accept.

[kelbyc]

I am having an issue now. I'm only getting the recovering email that says the device Returned from the down state. Any idea why I'm not getting the initial email saying the device is down?

[noname]

I'm only getting the recovering email that says the device Returned from the down state. Any idea why I'm not getting the initial email saying the device is down?

Though I'm not sure about it, what is stored in 'thold_failed_hosts' ?

Code: Select all

mysql> select * from settings where name = "thold_failed_hosts";

On trial, set empty value for it if some host-ids were stored.

[kelbyc]

I actually realized it is sending the correct message, it just has the wrong subject. I edited poller.php and now it actually says Down instead of recovering in the email. I'm getting both the down and recovering emails now correctly. Thanks for your help!

[noname]

I actually realized it is sending the correct message, it just has the wrong subject. I edited poller.php and now it actually says Down instead of recovering in the email. I'm getting both the down and recovering emails now correctly.

Hmm, it looks like this case, but anyway congratulations!
- http://forums.cacti.net/viewtopic.php?f=17&t=41838

[leolovemary]

Any idea why I'm not getting the initial email saying the device is down?

[TheWitness]

Make sure you've first tested the e-mails settings under Console->Settings->Mail/DNS

TheWitness