Timeouts - HDD vs SSD Performance Impact

[Olorin113]

Hello,

I'm experiencing timeout issues with kafka_exporter during Prometheus scraping.

I've run tests on both HDD and SSD, and I've noticed that the disk type seems to have a significant impact on performance.

My configuration
Environment:

System: Linux
Tests performed: HDD vs SSD
Prometheus configuration:

scrape_configs:

• job_name: 'kafka'
  scrape_interval: 30s
  scrape_timeout: 5s
  static_configs:
  • targets: ['kafka-exporter:9308']

kafka_exporter configuration:

kafka_exporter
--kafka.server=kafka-broker:9092

Observations
On HDD: Frequent timeouts, degraded performance
On SSD: Noticeable performance improvement

Question
Has anyone performed similar tests (HDD vs SSD) with kafka_exporter?

Did you observe similar differences? What optimizations have you applied to address performance issues on HDD?

Thanks in advance for your feedback! 🙏

Related issues:

Issue #513: Intermittent abscence for metrics for the kafkaconsumers_group
Issue #509:Metrics are not flowing continously for consumer group lag.
Recurring problems with a clusters (20 topics, 200 partitions)
Clean and straight to the point!