Full-stack observability and application monitoring with
summary
Concrete recipe for observability: use three pillars: logs metrics traces as primary signals, send traces to X‑Ray or OpenTelemetry, run collection with AWS Distro for OpenTelemetry (Collector, Lambda layer, or ADOT Operator on EKS), and visualize with Managed Prometheus workspaces and Amazon Managed Grafana. Claim: keep page loads <3s and purchase success high, use Synthetics, RUM, Contributor Insights, Metrics Insights and anomaly detection to cut MTTR. Prometheus data stores to S3, ingestion/query via SigV4, and hybrid on‑prem agents scale without reworking app code.
AI-written summary · about
chapters (19)
- 0:00Intro
- 0:35Agenda What observability is
- 1:38What is observability?
- 3:26Observability vs. monitoring
- 5:01The three pillars of observability
- 9:54AWS-native observability stack
- 12:28Full-stack observability strategies
- 22:38Inside-out
- 26:43A well-developed, full-stack example
- 27:50What about hybrid, distributed, or on-premises workloads?
- 29:51Open source managed services
- 32:39Amazon Managed Service for Prometheus
- 36:52Typical metric collection setup
- 38:27Amazon Managed Grafana
- 40:55Flexible data source options and visualization
- 44:57AWS Distro for OpenTelemetry
- 47:05OpenTelemetry Kubernetes operator
- 49:20Flexible options
- 52:44Optimize IT for the cloud to speed migration
related talks
24mObservability with AWS
37mBuilding Modern Observability Architecture With OpenSearch, OS Ingest...
35mWhat’s new with observability at AWS?
46mObservability best practices for hybrid environments
36mObservability at Scale With Monitoring as Code...
01h 02mDesigning an open-source observability stack for Kubernetes
16mBuilding Observability into Cloud-Native Applications
43mImplementing an effective observability strategy