System context

rasdaemon is a privileged host daemon that reads Linux kernel RAS trace events, decodes them, and reports them to system logging. Events can also be stored in a persistent database, either locally or remotely, using a backend. Currently available database backends are SQLite, MySQL/MariaDB, or PostgreSQL.

Optional consumers can apply containment or isolation policies and/or report selected events to other platform services.

The principal stakeholders are:

Platform operator

Installs, configures, monitors, and upgrades the daemon. The operator needs predictable startup, actionable diagnostics, safe defaults, and recovery procedures.

Reliability engineer

Uses event history to diagnose hardware faults and trends. This user needs accurate decoding, timestamps, hardware location information, and a stable query interface.

Platform integrator or distributor

Selects features, database backends, architecture support, service policy, and packaging. This user needs build-time isolation, compatibility information, and a verifiable release process.

Developer or maintainer

Adds trace-event decoders, actions, and backends. This user needs explicit interfaces, ownership rules, focused tests, and useful failure diagnostics.

Security or safety assessor

Evaluates privilege, external inputs, active mitigation features, data handling, and failure modes. This user needs a threat model, traceable controls, and evidence from testing.

External entities

May receive reports, such as ABRT reports, to support software quality monitoring, capability management, and planned system or component decommissioning.

The Linux kernel trace-event ABI and optional database or BMC interfaces are external dependencies. Statistical failure prediction, fleet analytics, kernel error injection, and operation of downstream deployments remain outside the daemon’s current scope; see Overview and Testing.