AWS has launched CloudWatch Omni, an observability workspace organised around applications rather than a collection of dashboards. It combines logs, metrics, traces, dependency mapping and investigations assisted by Amazon DevOps Agent.

The short answer

QuestionAnswer
Does it replace CloudWatch?No. Omni reuses existing signals and alarms.
Do teams need AWS Console access?No. They use a dedicated URL with enterprise SSO.
Which standards are supported?OpenTelemetry and an OTLP endpoint for external environments.

Applications before infrastructure

Omni discovers services from telemetry and AWS Config, then builds a topology that changes with deployments. Teams declare availability, latency and error objectives instead of maintaining every view manually. This organisation can reduce gaps between infrastructure components.

A shared investigation

When an alarm fires, a session collects correlated signals, recent changes and the history of decisions. An SRE can invite the owning team without rebuilding context in a separate chat thread. The goal is to preserve the timeline through mitigation and post-incident review.

DevOps Agent must remain an assistant

The agent searches for correlations, proposes causes and suggests mitigation steps. Its conclusions still depend on trace quality and an accurate service map. Teams should confirm causality before a rollback or automated action, especially during a complex incident.

Plan cost and governance

Before broad deployment, organisations need to define Spaces, SSO rights, retention and OpenTelemetry volumes. Centralising more signals can improve investigation while increasing ingestion and exposing sensitive data. A pilot should measure both time saved and the resulting bill.