PagerDuty launches PD Reliability Platform with Paige PagerDuty launches PD Reliability Platform with Paige

One platform, different release stages

The platform replaces the Operations Cloud name and offers purchasing across events, actions and seats. Customers can buy flexible credits and allocate spending between human users and agent actions. The launch offer includes additional AI Actions capacity for the first subscription year for qualifying Essential, Plus and Ultimate customers who purchase before January 15, 2027.

Approval stays in the incident path

Paige suggests resolutions using similar previous incidents and can run specific available runbooks, including service restarts or Kubernetes remediation. The explainer distinguishes approved resolutions from planned autonomy. Teams choose whether to investigate further, apply a recommendation or take another path. Deep Investigation and SRE Workbench sit in PD Labs as previews. The first tests competing root-cause explanations; the second gives responders and the agent a shared investigation space.

Evidence is more than a matching timestamp

PagerDuty’s engineering account describes a diagnosis as two steps: locating the failing component and explaining what failed inside it. The investigation records both direct and circumstantial evidence. A configuration change that removed a required setting directly supports a failure mechanism; simultaneous alerts only suggest a place to investigate. It also keeps rejected explanations and counter-evidence. Investigations update as logs and notes arrive. Responders can redirect an active investigation.

Connecting the services behind an alert

The same engineering post describes topology collection from source code, deployment configuration and infrastructure definitions. Extractors record dependencies with file and line references, then a central engine resolves links across repositories. The collector can run through a GitHub integration or inside a customer’s own continuous-integration system. This supplies relationships that an alerting service alone cannot explain, such as its dependence on another team’s queue or database.

An established operations-agent market

PagerDuty already described its agent-assisted Slack incident workflow on September 16. The new platform also enters a market with AWS DevOps Agent, which became generally available on March 31. AWS describes investigations across its cloud, other clouds and on-premises environments, correlating telemetry with code and deployment data. Its service connects to observability tools, runbooks, repositories and delivery pipelines. AWS also links monthly usage credits to the previous month’s AWS support plan bill: 100% for Unified Operations, 75% for Enterprise Support and 30% for Business Support+, under its published program. The AWS release also describes custom agent skills to extend capabilities, plus custom charts and reports for operational insights. AWS also directs preview customers to migration documentation for access to its new capabilities.