Elastic Observability Labs
9 July 2026
From five dashboards to one prompt: how we built an APM health monitor with Elastic Agent Builder
Five ES|QL tools score latency, errors, throughput and dependencies to find the root cause, so you don't dashboard-hop during an APM incident.
Featured
Use Elasticsearch as a Drop-In Prometheus Backend for Grafana
Use Elasticsearch as a Prometheus backend for Grafana dashboards, autocomplete, Metrics Drilldown, and alerting without changing PromQL workflows.
From alert to root cause in seconds: AI-powered observability with Elastic Agent Builder and Workflows
Elastic Agent Builder and Workflows replace dashboard hunting: one question surfaces the root cause, correlates metrics across weeks, and calculates business impact; then the workflow files the ticket.
3 signals, 2 env vars, 0 collectors: OpenTelemetry with Python and Elastic's Managed OTLP Endpoint
Instrument a Flask API with OpenTelemetry and ship traces, metrics, and logs to Elastic Cloud using just 2 environment variables, no collector needed.
Contextual AI: Stop pinging the SRE: three MCP tools that turn Elastic Agent Builder into your team's runbook
Build three MCP tools in Elastic Agent Builder that read endpoint health, recent deploys and SLO burn rate directly in your editor. Encode your platform team's runbook once; every developer gets self-serve production context without pinging an SRE.
SNMP Topology Data in Kibana: Collection to Canvas
The Network Topology plugin for Kibana provides a ready-to-deploy Logstash pipeline, a structured schema, and a topology view that shows what's connected to what.
Configure downsampling directly in Elastic Streams, no more JSON editing needed
Configure downsampling in Elastic Streams alongside retention and tiers, with a live preview and validation. No more editing ILM or lifecycle JSON.
Self-Driving Observability: From Stacktraces to Profiling-Derived Metrics
Profiling-derived metrics turn raw stacktraces into time-series KPIs, unlock continuous profiling for every user and lay the foundation for an observability system that detects, investigates, and acts on its own.
Don't leave metrics on the table: query them with the ES|QL TS command
Recalibrate your mental model for time series queries: learn why FROM can produce inaccurate results for metrics, how TS fixes that, and when to use each command.
Bringing Fire to Elasticsearch: Adding Native Prometheus API Support
Query Elasticsearch directly from Prometheus-compatible clients via native PromQL, discovery, and metadata endpoints. Send data to Elasticsearch with Prometheus Remote Write.
From averages to any percentile: Elasticsearch ships native exponential histogram support in ES|QL
Query any percentile at any time. Elasticsearch natively stores OTel exponential histograms and lets you analyze distributions in ES|QL without fixed buckets or lossy conversions.
ES|QL queries for debugging LLM latency, cost and GPU saturation
Learn how to investigate LLM latency, token cost and GPU saturation using ES|QL against OpenTelemetry traces and get a root cause, not just a symptom.
30x faster than Prometheus: how we rebuilt Elasticsearch as a leading columnar metrics datastore
Elasticsearch now stores OTel metrics at 3.75 bytes per data point and queries them up to 30x faster than Prometheus. Here's how we rebuilt TSDS and ES|QL.