Skip to content

Glossary

A configured link between Neubird and an external service (cloud provider, monitoring tool, incident management platform, etc.) that enables Neubird to query telemetry data. Connections use read-only credentials and can be managed via the Platform or MCP Server. See Connections overview.

NeuBird’s core discipline of providing an LLM with the right information, in the right format, at the right time. Context engineering — combined with data virtualization — is what enables Neubird to outperform competitors using more powerful foundation models.

A unified query schema across all observability tools. A single investigation can correlate data from Datadog metrics with Splunk logs with AWS CloudTrail events, without needing tool-specific query languages.

A structured, six-phase infrastructure health check available in the Desktop CLI via the /health command. Covers topology, metrics, logs, traces, recent changes, and cost signals. See Health assessments.

An interactive or automated analysis session where Neubird queries telemetry data, follows evidence, and produces root cause analysis. Investigations can be started manually or triggered automatically by alerts.

Custom guidelines attached to a project that tell Neubird how to approach investigations for specific services or environments. Instructions can include investigation strategies, escalation rules, and environment-specific context. See Project instructions.

Knowledge that Neubird accumulates from investigations to improve future analysis. In Desktop, this includes SQL rules and table mappings stored locally. In Platform, past investigation patterns are stored in a vector database. See Learnings & memory.

Metrics, Events, Logs, Traces — plus configuration, code, and infrastructure definitions. Neubird’s telemetry access goes beyond traditional MELT to include Kubernetes manifests, deployment scripts, and source code.

A container in Neubird that groups services, resources, and connections for investigation. Projects scope what data Neubird can access during an investigation session. See Projects.

The final output of a Neubird investigation. An RCA identifies the specific change or event that caused an issue, when it took effect, the causal mechanism, and recommended remediation actions — all backed by evidence.

A documented procedure for investigating or remediating specific types of incidents. Runbooks can be attached to projects to guide Neubird’s investigation approach. See System runbooks.

An active investigation interaction within a project. Sessions maintain conversation context so you can ask follow-up questions. Sessions are saved and can be reviewed later. See Sessions overview.

Code generated by Neubird’s fine-tuned LLM that fetches and processes data from your integrations. Telemetry programs are written in NeuBird’s RAEL language and execute in isolated, ephemeral environments.


A unique identifier used to authenticate API requests between Neubird and external services like Datadog, New Relic, or PagerDuty.

An AWS service that records account activity across your AWS infrastructure, providing governance, compliance, and operational auditing data.

AWS’s monitoring and observability service that collects metrics, logs, and alarms from AWS resources and applications.

A string passed when making an AWS request to assume a role. Ensures that only trusted entities (like Neubird) can access the role.

AWS service that manages secure access to AWS resources by defining user and service permissions and policies.

An event that disrupts normal operations, requiring investigation and resolution. Typically tracked in platforms like PagerDuty, ServiceNow, or incident.io.

Records of system or application activities used for troubleshooting, monitoring, and auditing.

Quantitative measurements of system performance and health — CPU usage, memory consumption, response time, error rates, etc.

A globally unique AWS identifier that specifies an IAM role. Used for cross-account access, allowing Neubird to assume a role with defined permissions in your account.

The automated collection and transmission of data from systems for monitoring and analysis. Includes metrics, events, logs, and traces.

Distributed tracing data that follows requests across service boundaries, showing the full path and timing of a request through a distributed system.