Probe Offline Monitoring Mode
A monitoring probe can keep monitoring while it is disconnected from the NetCrunch Server, buffering results locally and synchronizing them once the connection returns.
Atlas SettingsSitesMonitoring
When a probe loses its connection to the NetCrunch Server, Offline Monitoring Mode lets it continue working on its own. The probe keeps polling its nodes and keeps processing triggers, and the results are buffered on the probe machine.
Once the connection is restored, the probe synchronizes the buffered data and returns to normal real-time communication. Monitoring continuity is preserved across the outage.
This makes monitoring resilient in locations with unstable or intermittent connectivity, where a link that drops for an hour would otherwise leave a gap in the data.
Offline Monitoring Mode is a property of the site, not of an individual probe. It cannot be used on a site that has load balancing enabled, because in that mode nodes are not permanently assigned to a particular probe. See Distributed Monitoring.
Enabling Offline Monitoring
Offline monitoring is enabled in the monitoring properties of a network site, where you also set how long the probe may keep collecting data while disconnected. The default is 24 hours.
The limit protects the probe machine from filling its disk during a long outage. When it is reached, the probe stops buffering new results.
offline-trends
Trends
All trend data collected while the probe was offline is synchronized with the server after the probe reconnects. Trends are continuous across the outage — you do not lose the period the probe spent disconnected.
offline-alerts
Alerts
What happens to an alert action depends on whether the probe was online when the alert opened and when it closed.
- Alert opened while online
- The alert action runs normally. If the alert is later closed while the probe is offline, the closing action runs once the connection to the server is restored.
- Alert opened and closed while offline
- No alert actions run. The reason is recorded in the alert action log.
- Alert opened while offline
- The alert action runs after the connection to the server is restored.
offline-exceptions
What Is Not Recorded
Some events are not generated for the period the probe spent offline, even where the underlying condition did occur during that time.
- Node Up/Down alerts
- Network Service Up/Down alerts
- Threshold alerts defined on Network Service counters — all other threshold alerts are recorded in the alerts history
- Bandwidth Sensor trends
- Conditional alerts
- Receivers (Syslog, SNMP Trap, Web Messages) — messages are not collected while the probe is offline. Only alerts defined on these receivers are written to the alerts history.
- Configuration sensors (Device Config, Windows Configuration and similar) — these are not executed while the probe is offline, and run again after it reconnects.
Plan around this list when a site is expected to spend long periods disconnected. Availability reporting for that site will not account for the offline window, because Node Up/Down events are among the events that are not generated.
Configuration Changes During an Outage
Nodes, sensors and services can be added or removed on the server while the probe is offline. The probe picks up those changes when it reconnects and applies them before resuming normal operation.