Extending OpenSearch Dashboards: Building a Unified Obser...
summary
Freshworks turns OpenSearch Dashboards into a full observability stack handling petabytes of logs plus traces, metrics, frontend and synthetic checks by building an adapter service with three modules: request interceptor/parser, request converter and response converter to query multi-cluster and multi‑data‑source backends. Migration from Elasticsearch to OpenSearch (license and searchable snapshots) enables this. Features include a service map, trace gang‑chart, profile diff view and alerts heatmap; result is lower operational cost, reduced alert fatigue and faster MTTR after a two‑year rollout.
AI-written summary · about
chapters (14)
- 0:00Unified observability intro
- 3:11Legacy observability challenges
- 5:22Why OpenSearch Dashboards
- 7:10Extending for traces
- 8:37Service maps and insights
- 9:18Alerts performance analysis
- 10:40Reducing alert fatigue
- 11:22Profiling and optimizations
- 12:03Front-end observability
- 12:44Synthetic monitoring
- 13:47Architecture challenges
- 15:52Adapter service intro
- 18:20Unified platform benefits
- 19:42Conclusion and future
related talks
06mSponsored Keynote: OpenSearch - See Everything: Open Observability for Agentic AI
25mUnifying Observability: Weaving Prometheus, Jaeger, and Open Source Together to Win
11mOpenTelemetry + OpenSearch Integration
20mMeasuring Data with OpenSearch to Better Serve the Community
37mBuilding Modern Observability Architecture With OpenSearch, OS Ingest...
21mKubernetes Under the Magnifying Glass: Observability With OpenSearch
31mFluent Bit & OpenSearch: The Power Couple of Open Source Observability
06mWhy OpenSearch?