Back to vendors
N

New Relic

Also known as: newrelic.com, New Relic Autopilot, New Relic SRE Agent, New Relic AI, New Relic Agentic Platform, New Relic Ground Truth, New Relic Workflow Automation

Visit site
Entry priceFree tier; Standard full platform users $10 for the first, then $99 eachFull pricing detail

New Relic sells an observability platform for applications, infrastructure, logs and AI systems, and runs agents on top of it. Autopilot, its SRE agent, starts work when an alert fires, triages the incident, finds likely root causes using runbooks and long term memory, and proposes fixes in New Relic or Slack. Workflow Automation executes rollbacks and restarts behind Slack approvals, the Agentic Platform lets teams build their own agents, and an MCP server feeds other agents New Relic data.

New Relic, founded in San Francisco in 2008 and taken private by Francisco Partners and TPG in a deal valued at about $6.5 billion that closed in November 2023, sells the Intelligent Observability Platform: application performance monitoring, infrastructure, logs, browser and mobile monitoring, synthetics, Kubernetes and AI monitoring on one telemetry database queried with NRQL. Domino's, Verizon, Toyota, Atlassian and William Hill are among its customers. Its agent work spans an SRE agent run by New Relic, a builder for customers' own agents, workflow automation that acts in cloud systems, and tools that feed other companies' agents.

New Relic Autopilot, the SRE agent and the default agent in the New Relic AI panel, starts analysis the moment an alert fires, triages the incident, separates noise with intelligent root cause analysis over the topology graph, identifies likely causes, measures user impact and scopes remediations, often before an engineer acknowledges the page, then helps write the post incident report.

It works in the New Relic interface, in Slack and Zoom triage rooms, and as a workflow action fired by a SEV1, a deploy or a schedule, grounds its conclusions in runbooks, retrospectives and documents connected through New Relic AI Knowledge, pulls code context from Jira and GitHub over MCP, and keeps long term memory of team knowledge that can be scoped and managed.

Workflow Automation, generally available since February 2026, carries out fixes such as restarting services, resizing EC2 instances and rolling back Lambda, API Gateway or AWS AppConfig changes, with approval steps that post a proposed action to Slack or Teams and wait for a person. The Agentic Platform, in preview, adds a drag and drop builder for SRE teams' own agents, an orchestration layer and dynamic runtime for multi agent work, MCP tool access, role based access, audit logs and an evaluation engine that keeps testing agents. Ground Truth and the New Relic MCP server give Claude Code, GitHub Copilot and custom agents dozens of read only tools over the same data.

Pricing is usage based: a free tier covers 100 GB of ingest a month and one full platform user, Standard full users cost $10 for the first and $99 for each more up to five, Pro full users $349 a month on annual commitment, core users $49, and data beyond the free allowance $0.40 per GB or $0.60 with Data Plus, with compute measured in Core Compute Units and AI features under Advanced Compute. Data can sit in US or EU data centers, New Relic maintains SOC 2, ISO 27001 and ISO 42001 among other programs, and Enterprise with Data Plus is eligible for FedRAMP Moderate and HIPAA. New Relic AI runs on Azure OpenAI models New Relic selects.

It fits engineering teams already sending telemetry to New Relic that want first response on incidents handled before people join, with fixes gated by approvals in Slack. It fits less well for teams that want to choose the model behind the agent, need agents that drive browsers, or want flat pricing, since agent investigations and MCP calls draw on usage meters.

Vendor details

Canonical URL

https://newrelic.com

Category

SRE / DevOps agent

Subcategory

Observability platform with Autopilot, an SRE agent that triages incidents and proposes fixes, workflow automation that executes remediations behind approvals, and an agent builder

Funding status

Private, owned by Francisco Partners and TPG since its take private closed in November 2023 in a deal valued at about $6.5 billion.

Company status

independent

Use cases & customers

Primary use cases

automated incident triage and root cause analysisapproved auto remediation and deployment rollbackcustom SRE agents built without codepost incident reporting and user impact measurementobservability data for outside coding and ops agents over MCP

Target customers

SRE and platform engineering teamsDevOps and on call engineersenterprises standardizing observabilityteams running AI applications and agents in production

Deployment options

SaaS with US or EU data centersFedRAMP Moderate eligible Enterprise with Data Plus

Integrations

Autopilot and Workflow Automation act through Slack, Zoom, Microsoft Teams, Jira, GitHub, PagerDuty, AWS (EC2, Lambda, API Gateway, Systems Manager, AppConfig), Azure and any HTTP API. The New Relic MCP server serves Claude Code, GitHub Copilot, Cursor, VS Code and other MCP clients through US, EU and JP endpoints, and the platform ingests telemetry from hundreds of languages, clouds and tools.

In practice

A checkout service's error rate alert fires at 2 a.m. Autopilot starts at once, traces the spike to a deploy 12 minutes earlier, measures the users affected and posts its findings in the incident Slack channel before the on call engineer has acknowledged the page.

After a Lambda release degrades latency, a Workflow Automation rollback posts the proposed change to Slack; the engineer reacts with approval and the workflow rolls the API Gateway integration back to the previous Lambda version.

A platform team uses Claude Code with the New Relic MCP server to check golden metrics and recent error groups for a service while writing a fix, without leaving the editor.

Agentic Index coverage score

12.0 / 14 capabilities · 86%

Integrations & Tool Calling Full

Workflow Automation acts directly in AWS, restarting or resizing EC2 instances and rolling back Lambda, API Gateway and AppConfig changes through Systems Manager, and reaches Azure, Slack, Jira, PagerDuty and any HTTP API from its action catalog. Autopilot pulls code context from Jira and GitHub over MCP connections and answers inside Slack and Zoom triage rooms where responders already work.

Sourcedocs.newrelic.com/whats-new/2026/02/whats-new-02-24-workflow-automationread 2026-11-18

Workflow Orchestration Full

Workflows are structured multistep processes with conditional logic, approvals and outside tool steps, built from templates or YAML, and the SRE agent can call a workflow or run as a step inside one. The Agentic Platform, in preview, adds a central orchestration layer and a dynamic runtime so custom agents can reason through multistep and multi agent work instead of a fixed script.

Sourcenewrelic.com/press-release/20260224read 2026-11-18

Knowledge Grounding & RAG Full

New Relic AI Knowledge connects Confluence retros, postmortems, runbooks and uploaded PDFs, CSVs and text files, refreshes them on a schedule, and Autopilot searches that content during each investigation. Typing a slash attaches a specific runbook or postmortem to a prompt, and the agent says when tool results came back truncated or empty so gaps in what it saw stay visible.

Sourcedocs.newrelic.com/docs/agentic-ai/sre-agent/use-sre-agentread 2026-11-18

Human Oversight & Guardrails Full

Workflow approval steps post a proposed action to Slack or Microsoft Teams and wait for a person to approve or reject it before a critical change runs, and teams choose which workflows run fully automatic. Autopilot keeps a human in the loop for certain decisions and offers configurable controls over its access and permissions before it takes on more automated tasks.

Sourcenewrelic.com/blog/observability/from-alert-fatigue-to-auto-remediation-new-relic-workflow-automationread 2026-11-18

Security, Identity & Governance Full

An organization admin must switch on MCP access in the feature control manager, and clients sign in with OAuth or a user API key scoped to that user's accounts. The Agentic Platform governs custom agents with role based access and audit logs, and workflows keep cloud credentials in secrets management. New Relic maintains FedRAMP, ISO 27001, ISO 42001, PCI DSS, SOC 1, SOC 2 and TISAX programs and HIPAA enabled capabilities, with newer services joining each program at its next audit, and Enterprise with Data Plus is eligible for FedRAMP Moderate and HIPAA.

Sourcedocs.newrelic.com/docs/agentic-ai/mcp/overviewread 2026-11-18

Observability & Auditability Full

Autopilot shows each step of an investigation with a checkmark and flags steps that ran into trouble, and the incident timeline lays out events before and after the failure. Custom agents carry audit logs, the platform records workflow runs with their logs, and teams monitor their own AI applications and agents with the same telemetry tools New Relic sells.

Sourcedocs.newrelic.com/docs/agentic-ai/sre-agent/use-sre-agentread 2026-11-18

Memory & State Persistence Full

Autopilot keeps long term memory that captures a team's operational knowledge and makes it available across teams, and that memory can be scoped and managed rather than left opaque. It sits beside the connected knowledge sources and the platform's own incident timeline, so a later investigation starts from what the team learned in earlier incidents.

Sourcenewrelic.com/press-release/20260623-2read 2026-11-18

Deployment & Data Residency Full

Accounts live in US or EU data centers, with EU storage priced at an extra $0.05 per GB, and the MCP server and its OAuth token service run separate US, EU and Japan endpoints so agents stay in the account's region. Enterprise with Data Plus is eligible for FedRAMP Moderate, while the platform itself runs only as a hosted service with no self hosted edition.

Sourcenewrelic.com/pricingread 2026-11-18

Prebuilt Agents / Templates / Packs Full

Autopilot arrives ready to run with a Kubernetes domain specialist, Kafka troubleshooting tools and cross stack root cause analysis, and the Agentic Platform ships prebuilt agents beside its builder. Workflow Automation offers templates for deployment rollback, EC2 scaling and incident enrichment, and the nr-skills-hub repository holds skills such as APM error investigation for outside agents.

Sourcenewrelic.com/press-release/20260623-2read 2026-11-18

Triggers & Channel Coverage Full

Autopilot starts its analysis the moment an alert fires, often before an engineer acknowledges the page, and runs as a workflow action fired by a SEV1, a deploy or a schedule. Workflows themselves start from firing alerts, metrics crossing thresholds or entities changing state, and people can also question the agent on demand in the AI panel or by mentioning it in Slack.

Sourcenewrelic.com/press-release/20260623-2read 2026-11-18

Model Flexibility & Routing Not documented

New Relic AI runs on large language models served through Azure OpenAI that New Relic chooses, and the SRE agent pairs those models with deterministic analytics such as causal graphs and topology scoring. Customers have no setting to pick a model, bring their own provider or route agent work to another model, so model choice stays entirely with New Relic.

Sourcenewrelic.com/blog/ai/protecting-data-privacy-and-security-with-generative-airead 2026-11-18

APIs / SDKs / MCP Extensibility Full

The NerdGraph API reads and changes the platform and is what workflows call to act, and the hosted MCP server at regional addresses accepts OAuth or API keys and lists its tools by tag, from execute_nrql_query to analyze_deployment_impact, though all of them only read. Ground Truth tools and the nr-skills-hub give Claude Code, GitHub Copilot and custom agents the same data.

Sourcedocs.newrelic.com/docs/agentic-ai/mcp/tool-referenceread 2026-11-18

Testing, Debugging & Optimization Full

A built in evaluation engine in the Agentic Platform tests agent performance continuously so custom agents earn trust before they act alone, and the platform is open to customers in preview. Teams also see each Autopilot step marked complete or flagged, and an agent states when truncated or empty tool results limited what it could check during an investigation.

Sourcenewrelic.com/press-release/20260224-1read 2026-11-18

Browser / Computer-use Not documented

Agents work through telemetry queries, APIs and workflow actions rather than by driving a browser or desktop. Synthetic monitors replay scripted browser checks against the customer's own sites, which agents can read as results, but that is scheduled monitoring rather than an agent operating an application screen on someone's behalf during an incident.

Sourcedocs.newrelic.com/docs/agentic-ai/mcp/tool-referenceread 2026-11-18

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

Free tier; Standard full platform users $10 for the first, then $99 each

Users by type plus data ingest per GB plus compute in Core Compute Units

Free tier

Included quota

100 GB of data ingest a month, one full platform user and unlimited basic users

What is public

The free tier covers 100 GB of ingest a month, one full platform user and unlimited basic users. Standard full users cost $10 for the first and $99 for each more up to five, Pro full users $349 a month on annual commitment or $418.80 monthly, and core users $49; Enterprise is quoted.

Billing mechanics

Charges combine users by type, ingest beyond the free 100 GB at $0.40 per GB or $0.60 with Data Plus, and compute measured in Core Compute Units, with New Relic AI under Advanced Compute.

Cost watchouts

Agent investigations and MCP tool calls draw on compute meters, and the analysis tools that generate reports or convert language to NRQL need the Advanced Compute add on.

Variable cost rationale

Bills rise with data ingested, compute consumed by queries and agent work, and the number of full platform users.

Additional watchouts

No per unit price appears for Core Compute Units or Advanced Compute, so agent and MCP usage is hard to estimate in advance.

Overage / add-ons

Data beyond 100 GB costs $0.40 per GB, or $0.60 with Data Plus, plus $0.05 per GB for EU storage; most MCP tools draw Core Compute Units and some draw Advanced Compute units

Sales call required

Mixed (some tiers require a call)

Free / trial

Free tier: 100 GB ingest a month, one full platform user, unlimited basic users

Lowest paid plan

Standard, $10 for the first full platform user and $99 for each more up to five, plus usage

Commercial notes

Data Plus adds up to 90 days of extra retention, log obfuscation and streaming export, and Enterprise with Data Plus is eligible for FedRAMP Moderate and HIPAA.

Missing data

New Relic gives no price per compute unit, no separate price for Autopilot or the Agentic Platform, and no Enterprise list price.

Agentic Index verified 2026-11-18

Alternatives to New Relic

Other products that do the work New Relic does, and what each one is.

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.