Amazon Web Services (AWS) Launches 'Amazon CloudWatch Omni'

[Financial News] Amazon Web Services (AWS) launched 'Amazon CloudWatch Omni' within Amazon CloudWatch. The service provides observability for applications and artificial intelligence (AI) agents in a single environment. It also automatically discovers and maps services and dependencies, while enabling automated alerts that adapt to application changes.
According to industry sources on the 2nd, CloudWatch Omni is an AI-powered observability experience for applications and AI agents operated by users. It provides agent and application observability in a single environment. With CloudWatch Omni, engineering teams can collaborate using the same data and investigative context. Users access a dedicated organizational URL and sign in with their existing ID. This allows them to use CloudWatch Omni without access to the AWS Management Console.
Engineering teams generally spend a significant portion of their observability work maintaining dashboards and adjusting thresholds. When an incident occurs, they also spend considerable time switching between multiple tools to understand the situation. If a problem affects several teams, relevant context is not naturally passed to the next engineer and can be lost in Slack threads or screenshots. CloudWatch Omni changes this approach. It organizes observability around applications rather than individual signals and enables the entire team to work together in the same workspace.
CloudWatch Omni is built on OpenTelemetry. Telemetry already sent to CloudWatch appears in CloudWatch Omni without any additional configuration. Telemetry from other workloads instrumented with OpenTelemetry can also be sent to an OpenTelemetry Protocol (OTLP) endpoint.
When an incident occurs, CloudWatch Omni opens an investigation session with relevant context already included. A typical incident investigation proceeds as follows.
When an alert indicates a rising error rate in the order payment service, CloudWatch Omni opens an investigation session. The session displays the service topology along with related signals, including a deployment that took place 10 minutes earlier and increased latency in a downstream payment application programming interface (API). It also shows the DevOps agent's initial analysis.
The on-call site reliability engineer (SRE) confirms the correlation between the deployment and the incident, then opens the trace view to identify the endpoint where the error occurred. The engineer next checks whether the payment API's latency is related to capacity limits.
When the SRE escalates the investigation to the payments team, a payments engineer joins the same session. The engineer can view everything identified so far. The session also provides the DevOps agent's analysis of the correlation between a configuration change to the payment service provider's API gateway and the incident. After identifying the root cause, the engineer rolls back the change. The entire investigation history is recorded automatically, eliminating the need to prepare a separate incident report.
CloudWatch Omni organizes telemetry around applications rather than infrastructure components. Based on telemetry data and resource discovery through AWS Config, the system automatically discovers services and maps their dependencies. This allows users to view an application as a single connected system rather than a collection of separate resources.
[email protected] Jang Min-kwon Reporter