Glossary
Key terms and concepts used throughout the OpsTrails documentation.
- Change Failure Rate
- A DORA metric measuring the percentage of deployments that cause a failure in production. OpsTrails helps track this by recording every deployment and correlating them with incidents and metric changes via Impact Analysis.
- CloudEvents
- An open specification (v1.0) for describing event data in a common format. OpsTrails uses CloudEvents as the format for all events on the timeline, ensuring consistency and interoperability. See Core Concepts.
- Configuration Drift
- When the actual state of infrastructure diverges from its intended configuration. OpsTrails can track config-change events to maintain a record of configuration updates, making drift visible on the timeline.
- Data Load
- An event type representing the completion of an ETL job, data migration, or bulk data import. Recorded on the timeline with
type: data-load. See Data Pipeline Monitoring. - Deployment
- An event type representing code being deployed to an environment. The most common event type in OpsTrails, recorded with
type: deployment. See Deployment Tracking. - Event
- A discrete moment recorded on the OpsTrails timeline. Every event follows the CloudEvents 1.0 specification and includes required fields (type, source, time) and optional fields (subject, severity, version, data). See Core Concepts.
- Impact Analysis
- A feature that compares metrics in a before/after window around an event to assess whether a change (deployment, rollback, etc.) made things better or worse. Requires connected analytics providers. See Impact Analysis.
- MCP (Model Context Protocol)
- An open protocol that enables AI models to interact with external data sources and tools. OpsTrails exposes an MCP server that gives AI assistants read-only access to query events and metrics using natural language. See MCP Overview.
- MTTR (Mean Time to Recovery)
- A DORA metric measuring the average time from an incident start to resolution. OpsTrails reduces MTTR by enabling AI-powered “what changed?” queries that surface the root cause in seconds. See Incident Response.
- Operational Timeline
- The chronological history of all events in an OpsTrails organization. Events are added to the timeline when recorded via the API, SDK, CLI, or CI/CD integrations, and become queryable immediately.
- Release
- An event type representing a new version of a service being published. Distinguished from deployment (which is the act of deploying code to an environment). Recorded with
type: release. See Release Management. - Rollback
- An event type representing a deployment being reverted to a previous version. Recorded with
type: rollback. Rollback events are particularly useful during incident investigation as they mark the point when a fix was attempted. - Severity
- An optional event field indicating the impact level: LOW (routine operations), MINOR (notable changes), MAJOR (significant changes), or CRITICAL (high-impact events). Used to filter and prioritize events during investigation. See Core Concepts.
- Source
- A required event field identifying where the event originated, formatted as a URI (e.g.
//github.com/org/repo). Sources link events to specific services, pipelines, or infrastructure components. See Core Concepts. - Subject
- An optional event field providing context, most commonly the target environment (e.g. production, staging). Subjects are used to filter events and match them with analytics metrics during impact analysis. See Core Concepts.
- Type
- A required event field describing what kind of event occurred (e.g. deployment, rollback, release, data-load, incident, config-change). You can use any string value. See Core Concepts.