Driving down the cost of observability
summary
Observability routinely hits 10–30% of cloud spend; 99% of indexed logs are never searched, 95% of alerts come from five errors, and 99.9% of queries stop before seven days. Practical fixes: generate usage stats from Prometheus/OpenSearch query logs, define 3–5 use-case-driven data classes, convert noisy logs into Prometheus metrics, apply tail sampling to traces (OpenTelemetry/Jaeger), and index cautiously. Coralogix’s TC optimizer routes frequent/monitoring/compliance tiers to SSD, archive (S3/Athena) or no-index storage with ~65%/88% discounts. Ask vendors what cost-optimization tools they provide.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:31Why costs are high
- 3:05Industry cost facts
- 5:58Data usage analysis
- 7:24Rethinking data retention
- 8:59Strategies for efficiency
- 10:40Monitoring data usage
- 12:25Use case based management
- 14:20Architectural patterns
- 16:12Querying archived data
- 17:31Demand optimization
- 20:46Tail sampling techniques
- 22:14Open source recap
- 23:43Managing SAS vendors
- 25:33Evaluating vendor tools
- 28:20Coralogics solution
- 31:44Integrations and flexibility
- 33:34Conclusion
related talks
30mObservability is Too Expensive - Chris Cooney
27mObservability is Too Damn Expensive!
22mObservability is Too Expensive!
10mWhy your observability is costing you too much?
31mMy Logs Cost More Than My Services
15mThree key considerations for deploying best-of-breed observability
16mShrinking the Observability Bill
40mObservability that scales: Analyze data in stream and query remotely