The alert is triggered when the storage space utilization of a host exceeds the configured threshold (default: 90%).
Host { .labels.name }: The host storage usage ratio has exceeded { .threshold }.
The physical disk pool {.id} of the host { .labels.name } has exceeded { .threshold } storage usage.
When a host in the cluster has high storage space utilization, the space allocation policy tends to rebalance data across nodes. Virtual volumes on that node may have no local replicas, and the cluster may trigger cross-node data rebalancing tasks more frequently, consuming more bandwidth and physical disk resources.
When the storage space utilization of a single node reaches 100%, I/O requests that require new space allocation will fail, affecting data writes and data safety.
When multiple nodes in the cluster are under high storage load, if a failure occurs, the remaining space on other nodes may be insufficient to complete data recovery, posing a data safety risk.
The storage space utilization on the node has exceeded the configured threshold.
Space utilization on some nodes is significantly higher than on others
Possible causes include:
The cluster topology is not well-balanced, which is usually accompanied by a topology capacity imbalance alert. Refer to the corresponding alert description.
In a short period, a large number of virtual machines were cloned from a non-template virtual machine with high capacity, and data was written rapidly, causing data pressure to concentrate on a small number of nodes. In this case, use hot migration to distribute these virtual machines across other nodes, and apply write bandwidth throttling to relieve system pressure. Remove the throttling after the cluster capacity rebalancing is complete.
A large-capacity physical disk on the node has failed, causing storage space utilization to spike suddenly. Replace the failed physical disk. The alert will clear automatically after the node's normal capacity is restored.
Overall cluster storage space utilization is high
This will simultaneously trigger a cluster-level storage space utilization alert. Refer to the corresponding alert description.
