ElasticsearchMetricjvm.gc.collections.elapsed
GC elapsed timeSummary
Tracks the cumulative time spent in garbage collection across all GC cycles since JVM startup. Rising values indicate increasing GC overhead, which can signal heap pressure or inefficient memory allocation patterns. Sustained high GC duration (especially when combined with frequent collections) typically means the heap is undersized or the workload is creating excessive object churn, leading to degraded query and indexing performance.
Interface Metrics (8)
Datadog
OpenTelemetryDatadogDatadogDatadogDatadogDatadogDatadogRelated Insights (2)
When JVM heap usage stays above 85% for extended periods, garbage collection pauses increase dramatically, leading to node unresponsiveness, cluster state propagation failures, and potential split-brain scenarios.
Cassandra nodes running on the JVM can experience heap exhaustion where heap usage climbs to 80-90% and stays elevated without dropping after GC, leading to OutOfMemoryError or node instability. This manifests as timeout errors at the application layer before the node crashes.