intermittent Graph

1. You use incorrect cron time for poller-wrapper (run every 7 minutes):

*/7 *     * * *   root    /opt/observium/poller-wrapper.py  >> /dev/null 2>&1

we support only 5 min interval, must be:

*/5 *     * * *   root    /opt/observium/poller-wrapper.py  >> /dev/null 2>&1

2. Your devices (as I see) support v2c and v3 snmp auth, please use it instead v1. This is reason why you have polling issue.

3. Use rrdcached and prefer ssd hard disks as storage for mysql/rrd.

$config['snmp']['version'] = "v1"; // Default version to use

This is also a horrifically bad idea. You should never use SNMP v1 unless a device doesn't support SNMP v2c. If your devices are using SNMPv1, this would explain why they are very very slow to poll.

470 seconds seems pretty good considering you have a device which takes 420 seconds by itself. You should disable that device to see what the other devices take, then slowly decrease the poller threads to get it close to 300 seconds. Then re-enable that device.

You will always get gaps if you don't have a 5-minute poller period. This is how RRD works.

I try to set cron job back 5 minutes, poller setting ( attach )

Change back the cron job to 5 minutes and increase your pollers to i.e. 35 or 40 instances.

Hi Denis

this my setting and SC Polling info  ( attach )

Please share your settings and a screenshot of the Polling Information site.

hi Denis

I try to adjust. polling , but , i still have a itermintent graph.

any idea ?

Yes or trough the Webinterface it's also possible to adjust the number of polling instances.

thanks Denis.
 poller instances in cron?