Storage Backend Write Latency Bottleneck

critical

latencyUpdated Feb 7, 2026

Slow storage write operations block collector workers, causing span reception to slow and queues to back up, ultimately leading to dropped traces.

Sources

Performance Tuning Guide | Jaegerwww.jaegertracing.io

Frequently Asked Questions | Jaegerwww.jaegertracing.io

Jaeger Monitoring: Essential Metrics and Alerting for Production ...last9.io

Technologies:

JaegerSymptoms of this issue are visible in Jaeger metrics and logs

ElasticsearchThe root cause of this issue originates in Elasticsearch

CassandraThe root cause of this issue originates in Cassandra

Apache KafkaApache Kafka metrics correlate with this issue and help confirm diagnosis

How to detect:

Monitor storage backend write latency histograms. For Elasticsearch, watch indices.indexing.index_time_in_millis. For Cassandra, monitor write_latency_99th_percentile. Rising latency over time indicates storage can't keep up.

Recommended action:

Scale storage backend vertically or horizontally. For Elasticsearch, check disk I/O, shard allocation, and indexing throughput. For Cassandra, monitor pending compactions and tombstone accumulation. Consider using Kafka as an intermediate buffer.