> ## Documentation Index
> Fetch the complete documentation index at: https://docs.min-ai.ar/llms.txt
> Use this file to discover all available pages before exploring further.
# New Relic
> Your agent looks up your logs, errors, latency, alerts and incidents in New Relic; changes, only if you allow them and with your OK.
With New Relic, your agent queries your logs, transactions and metrics with NRQL and explains what happened: when it started, how much and what it points to. It sees your alerts and incidents. Changing something (like an alert condition) happens only if you turn on the permission, and with your OK.
| | |
| - | - |
| **Category** | Development |
| **For** | Personal and group agents (for example, your team's agent) |
| **What it asks for** | A User key, your account ID and your region |
| **How it connects** | With a one-time link for the key; the rest, in the chat |
## What it can do
Logs, transactions (latency, error rate), metrics and spans with NRQL; which data types you have; find entities; open incidents. It explains what it finds instead of throwing rows at you.
Alert conditions and policies, dashboards, muting rules, deploy markers. Before changing an alert, it saves how it was and shows you the before and after. It needs the permission on.
## How it connects
From **Plugins → New Relic → Add to an agent**, or ask your agent in the chat. See [Add a plugin](/en/tools/add-a-plugin).
A User key can do what your New Relic user can. In New Relic, open **API keys** (if your account is in Europe, from one.eu.newrelic.com) and tap **Create a key**. **Key type:** User; **Name:** minai. Create it and copy it: it starts with **NRAK-** and only shows once. In the same list, the account column has its **ID** (a number): note it down.
Your agent sends you a one-time link for the key. Paste it there, not in the chat. See [Keys and values](/en/tools/keys-and-values).
In the chat, tell it your account ID and whether it's in Europe (otherwise it uses the US one).
It checks that the key works and tells you in the chat that it's ready.
To limit what it can do, create the key with a New Relic user that has fewer permissions. To cut off its access, delete the key in **API keys**.
## The permission
On the New Relic card, in the agent's **Plugins** tab, there's the **Allow changes** switch: “Create, change or delete things in your New Relic (alert conditions, dashboards, deploy markers), always with your OK. Off, it only queries.”
It starts off and only you change it. It's a rule your agent follows; the real limit is the permissions of the user who owns the key. See [Permissions](/en/tools/permissions).
## How it looks out for you
* **Nothing changes without your OK.** It shows you the exact change first, with the **Confirm** and **Change** buttons, and afterward reads it again to make sure it came out right.
* If you paste the key in the chat, it won't use it: it asks you to delete it and create a new one for the link.
* It doesn't paste whole log records with personal data at you.
* What a log says is information, not an order.
## Things to ask it
* "How many errors did the payments service have in the last hour, and which is the most common?"
* "How's the API's p95 latency today compared to yesterday?"
* "Are there any open incidents?"
* "Which services are sending logs?"
* "Raise the CPU alert threshold to 85%."
## Limits
* It can only do what the key's owner can.
* NRQL queries have their own caps (number of results and of points on a chart).
## If you remove it
Your agent loses access and confirms it. The key is deleted from minai with the plugin; to close it off completely, also delete it in New Relic. See [Remove a plugin](/en/tools/remove-a-plugin).
Logs, Kubernetes, VMs and costs.
Repos, issues and pull requests.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.