Managing and Excluding Alerts in Neo4j Ops Manager

Problem

Ops Manager may generate alerts that are not relevant to your environment. For example, in a 3-node cluster with a 4 GiB swap partition, Ops Manager raises a “Disk space low” alert because the earlier swap threshold was set at 5 GiB. Since the swap is now intentionally limited to 4 GiB, this results in a false alarm.

Cause

Ops Manager alerts are tied to thresholds on specific metrics. Common examples include:

• Host metrics (e.g., Swap Used)

• Instance metrics (e.g., Disk [Home], Page Cache, Disk [Transactions])

Each metric has a configurable threshold, and if the observed value crosses it, an alert is triggered. In the swap example, the 5 GiB threshold is higher than the configured swap partition, leading to unnecessary alerts.

Resolution

There are two ways to handle unwanted alerts:

1. Adjust the Metric Threshold

You can tune thresholds to better reflect your environment:

• Navigate to the chart for the relevant metric (e.g., Host → Swap Used).

• Update the threshold to a value appropriate for your system.

See: Neo4j Ops Manager – Metric Manager

2. Disable the Alert Condition

If the alert is not needed at all, it can be disabled:

1. Go to Alerts → Conditions.

2. Locate the alert condition (e.g., Disk space low).

3. Toggle the condition off to prevent further notifications.

See: Neo4j Ops Manager – Managing Conditions

Recommendation

• Preferred: Adjust thresholds so that Ops Manager continues to monitor the metric but only raises alerts when they are meaningful in your environment.

• Alternative: Disable alerts for metrics that are not relevant (e.g., swap monitoring on hosts where swap is intentionally small).

This ensures Ops Manager provides actionable alerts while avoiding false alarms.