You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi All. I know this has been raised before but I am going to try to put some different view on it.
If you have multiple collectors and want to use a centralized database like MySQL all the collectors see the same data. This is good for all tables apart from the devices-related tables.
The idea behind my request is that all collectors will have all the OIDs and measurements and filters etc… But each collector will have its own list of devices to poll. These three tables look like the only ones that contain device-specific information.
snmp_dev_filters
snmp_dev_m_groups
snmp_device_cfg
If we can add for example the instanceID to the device then each collector will only add the devices that are intended for it.
The workaround we are testing is for now to have all devices loaded to all collectors. We then use an orchestrator to set the active/no active on each collector in the runtime config. The problem is that whenever we need to reload the config we then need the orchestrator to do a rebalancing or just a re-config.
Regards
Elad
The text was updated successfully, but these errors were encountered:
Hi All. I know this has been raised before but I am going to try to put some different view on it.
If you have multiple collectors and want to use a centralized database like MySQL all the collectors see the same data. This is good for all tables apart from the devices-related tables.
The idea behind my request is that all collectors will have all the OIDs and measurements and filters etc… But each collector will have its own list of devices to poll. These three tables look like the only ones that contain device-specific information.
snmp_dev_filters
snmp_dev_m_groups
snmp_device_cfg
If we can add for example the instanceID to the device then each collector will only add the devices that are intended for it.
The workaround we are testing is for now to have all devices loaded to all collectors. We then use an orchestrator to set the active/no active on each collector in the runtime config. The problem is that whenever we need to reload the config we then need the orchestrator to do a rebalancing or just a re-config.
Regards
Elad
The text was updated successfully, but these errors were encountered: