New Relic's infrastructure agent collects and displays data using six primary events, each with associated attributes that represent assorted metrics and metadata.
Understanding infrastructure data can help you:
- Better understand our infrastructure monitoring UI.
- Manage your infrastructure data.
- Run better queries of your data.
- Set up better monitoring solutions using custom attributes.
Sugerencia
Check out our troubleshooting tutorial for your infrastructure host data. The tutorial series walks you through how to find data in the infrastructure UI to resolve an incident and make a resource decision about your hosts.
Infrastructure events
The following are events reported by default by the infrastructure agent and some infrastructure integrations.
The attributes attached to these events are the metadata and metrics used to create our infrastructure UI visualizations. You can also create custom queries and charts of this data.
If you're using integrations, see that integration's doc for more on reported data. For common AWS attributes, see AWS data.
Select an event name in the following table to see its attributes.
Event | Description |
---|---|
| |
ImportanteProcess metrics are not reported by default. To report this data, enable process metrics. | |
ImportanteIf your server uses disks with file systems other than the supported file systems in the following table, | |
| |
| |
|
To learn about infrastructure integration data, see the documentation for a specific integration.
If an AWS integration is enabled, your infrastructure events may also have AWS attributes attached.
Query infrastructure data
You can query your infrastructure data to troubleshoot a problem or create a chart, or to understand what data is available. For example, to see what data is attached to ProcessSample
, you would run this NRQL query:
SELECT * FROM ProcessSample
Metrics from the New Relic infrastructure agent (SystemSample
, StorageSample
, NetworkSample
, and ProcessSample
) get aggregated into buckets to improve the query performance over longer periods of time.
NRQL queries spanning less than one hour directly access the raw data. For queries exceeding one hour, data is retrieved from the following aggregated buckets:
- 1 minute (scopes of 1 hour to 6 hours)
- 10 minutes (scopes of 6 hours to 3 days)
- 1 hour (scopes of 3 days to 14 days)
- 3 hours (scopes of 14+ days)
You can also query infrastructure using dimensional metrics.
Importante
Historically, these aggregated event buckets only contained one average value per metric. During October and November 2024, we're gradually migrating the aggregation to use the gauge format instead. This allows you to differentiate between max()
, min()
, and average()
values.
To check if your data is using the gauge format, run a query like this:
FROM SystemSample SELECT cpuPercent SINCE 90 minutes ago
If the result shows "type": "gauge"
, your data is using the new format.
Manage data
For tips on managing data ingest and reporting, see Manage infrastructure data.
Add custom attributes
You can create custom attributes in the infrastructure agent's YAML file. Use this metadata to:
- Filter your entities in the entity filter bar
- Populate the Group by menu
- Annotate your infrastructure data
Common Amazon EC2 attributes
If you use our Amazon Elastic Compute Cloud (EC2) integration, we report data from your Amazon EC2 instances. Amazon EC2-related attributes are common attributes present for any events reported from your EC2 instances. These attributes and their values are subject to change if Amazon changes the data they expose.
A sub-set of these attributes are collected from the infrastructure agent when installed in the EC2 instances:
awsAccountId
awsRegion
awsAvailabilityZone
ec2AmiId
ec2InstanceType