On-Premise Poller Scalability and Capacity Guidelines
As your organization grows, you can scale your monitoring environment by adding On-Premise Pollers and distributing monitoring workloads across them. This article explains how On-Premise Poller capacity scales, how On-Premise Poller Groups distribute workloads, the capacity considerations for network monitoring, and how monitoring continues during communication outages.
On-Premise Poller
An On-Premise Poller collects monitoring data from resources within your private network and sends the data to Site24x7. You can deploy multiple On-Premise Pollers when your monitoring environment requires additional capacity or spans multiple network locations.
The capacity of an On-Premise Poller depends on factors such as:
• Number and type of monitors
• Polling interval
• Amount of data collected
• Monitoring modules in use
• CPU and memory available to the On-Premise Poller
There is no single monitor-count limit that applies to all monitor types. As the monitoring workload increases, you can add more On-Premise Pollers and distribute the workload across them.
When to add another On-Premise Poller
Consider adding another On-Premise Poller when:
• The monitoring workload increases significantly.
• The existing On-Premise Poller approaches its available system resources.
• Resources are distributed across multiple network locations or segments.
• You need to separate monitoring workloads.
• You need a high-availability setup for critical monitoring environments.
On-Premise Poller Group
An On-Premise Poller Group is a collection of On-Premise Pollers that helps distribute supported monitoring workloads. A group can contain multiple On-Premise Pollers and can include primary and standby On-Premise Pollers for high availability.
Distribute monitoring workloads
On-Premise Poller Groups provide visibility into the workload and resource utilization of the On-Premise Pollers in the group. You can redistribute associated monitors when required to balance the workload.
As your environment grows, you can add On-Premise Pollers to the group and distribute supported monitoring workloads across them.
Provide high availability
An On-Premise Poller Group can include primary and secondary On-Premise Pollers to provide monitoring continuity if a primary On-Premise Poller becomes unavailable.
When the primary On-Premise Poller is unavailable, the standby On-Premise Poller takes over the monitoring resources assigned to it. The secondary On-Premise Poller must have connectivity to the resources it is expected to monitor.
For supported monitoring types, Site24x7 checks the primary On-Premise Poller’s status across consecutive polls before initiating failover.
On-Premise Poller Group requirements and limitations
Before adding an On-Premise Poller to a group ensure that:
• The On-Premise Poller is not already associated with monitors.
• The On-Premise Poller is not part of another On-Premise Poller Group.
• The On-Premise Poller is not down.
• The On-Premise Poller is not associated with a user-defined location profile.
• The On-Premise Poller meets the required version and network prerequisites. Refer to the On-Premise Poller system requirements for details.
An existing primary and secondary pair configured for On-Premise Poller High Availability cannot be added to an On-Premise Poller Group.
For network monitoring, On-Premise Poller Groups support network availability and performance monitoring. They do not support network traffic monitoring or Network Configuration Manager (NCM).
Network monitoring scalability
Network monitoring has separate capacity considerations because the workload depends on the number of devices, interfaces, performance counters, and flows being collected.
Network device, interface, and performance counter capacity
A single On-Premise Poller can support up to the following network monitoring capacity at a 15-minute polling interval, when deployed with the recommended system specifications:
| Network resource | Capacity |
|---|---|
| Network devices | 1,500 |
| Interfaces | 10,000 |
| Performance counters | 10,000 |
These are network monitoring capacity guidelines, not a universal monitor limit. Actual capacity can vary based on the monitoring workload and configuration.
For larger environments, deploy additional On-Premise Pollers and distribute the monitoring workload.
Scaling large network environments
Site24x7 recommends using an On-Premise Poller Group when monitoring more than 1,000 network devices. On-Premise Poller Groups can distribute supported network availability and performance monitoring workloads across multiple On-Premise Pollers.
Note: Network traffic monitoring and NCM are not supported by Poller Groups. Plan these monitoring workloads separately.
Network traffic monitoring
Network traffic monitoring has its own monitoring workload and is not supported through On-Premise Poller Groups and has a separate capacity consideration from network device and interface monitoring.
Site24x7 supports flow technologies such as NetFlow, sFlow, J-Flow, IPFIX, cFlow, NetStream, and AppFlow, with a flow-processing capacity of approximately 20,000–40,000 flows per second.
This flow-processing capacity should not be combined with the network device, interface, or performance counter figures above.
Network Configuration Manager
NCM has its own monitoring workload and is not supported through On-Premise Poller Groups.
When scaling an environment that uses NCM, consider its requirements separately from network availability, performance, and traffic monitoring.
Monitoring during communication outages
The On-Premise Poller continues collecting monitoring data if communication with the Site24x7 server is temporarily unavailable. The collected data is stored locally—with no time limit on local storage during the outage, and synchronized with the Site24x7 server after connectivity is restored. Learn more about data synchronization after internet connectivity outages.
Deployment considerations
Use the following approach when planning a scalable On-Premise Poller deployment:
| Requirement | Approach |
|---|---|
| Growing monitoring workload | Add On-Premise Pollers and distribute workloads. |
| Resources across multiple network locations | Deploy On-Premise Pollers where they can access the resources they monitor. |
| Large network environment | Use network device, interface, and performance counter capacity as planning guidelines. |
| More than 1,000 network devices | Consider an On-Premise Poller Group for network availability and performance monitoring. |
| Monitoring continuity | Configure high availability using primary and standby On-Premise Pollers. |
| Network traffic monitoring | Plan flow-processing capacity separately. Plan Network traffic monitoring capacity separately from On-Premise Poller Group capacity. |
| NCM | Plan NCM capacity separately from On-Premise Poller Group capacity. |
The number of On-Premise Pollers required depends on your monitoring workload, polling intervals, resource types, system capacity, and network topology.
Frequently asked questions
1. How many monitors can an On-Premise Poller support?
There is no single monitor-count limit for all monitor types. Capacity depends on the monitoring workload, polling interval, monitor type, and available system resources.
2. Can I add more On-Premise Pollers as my environment grows?
Yes. You can deploy additional On-Premise Pollers at any time and distribute monitoring resources across them through On-Premise Poller Groups. Adding On-Premise Pollers increases monitoring capacity, and you can rebalance monitors between On-Premise Poller groups as your workload changes.
3. What is an On-Premise Poller Group?
An On-Premise Poller Group is a collection of On-Premise Pollers used to distribute supported monitoring workloads and provide high availability. A group can contain multiple On-Premise Pollers, including primary and standby On-Premise Pollers for monitoring continuity.
4. When should I use an On-Premise Poller Group?
For network monitoring, Site24x7 recommends using an On-Premise Poller Group when monitoring more than 1,000 network devices. On-Premise Poller Groups distribute supported network availability and performance monitoring workloads across multiple On-Premise Pollers.
5. Does an On-Premise Poller Group support all network monitoring types?
No. On-Premise Poller Groups support network availability and performance monitoring. They do not support network traffic monitoring or NCM. Plan traffic monitoring and NCM workloads separately.
6. What happens when a primary On-Premise Poller goes down?
If high availability is configured, the standby On-Premise Poller takes over the monitoring resources after the primary one is confimed to be unavailable. Site24x7 checks the primary On-Premise Poller’s status across consecutive polls before initiating failover.
7. Does monitoring stop if the On-Premise Poller loses communication with Site24x7?
No. The On-Premise Poller continues collecting monitoring data locally and synchronizes the collected data after connectivity is restored. For default monitor data, reports that repeatedly fail to sync for two days are removed from the On-Premise Poller’s disk, so it's important torestore connectivity as soon as possible during extended outages.
8. How is backlogged data synchronized after connectivity is restored?
The On-Premise Poller sends current monitoring data first while processing the backlog. Network monitoring data from the last 15 minutes is sent first, followed by the remaining backlog. General monitor data is processed from oldest to newest.