Infrastructure alerting examples

The Alert type field in infrastructure monitoring's Settings > Alerts page shows what options you can select to create infrastructure alert conditions. You can also create alert conditions from any infrastructure chart by selecting the ellipses [ellipses icon] icon and then Create alert.

Examples: Infrastructure pages

Here are some examples of how to create alert conditions within the context of the Infrastructure monitoring UI page you are currently viewing. To create an alerts condition from any chart, select the ellipses [ellipses icon] icon and then Create alert. New Relic will automatically select the appropriate Alert type.

Example Problem and solution
High CPU usage

Problem:

Your Ops team monitors a filtered set of host clusters in your eastern region and notices that the CPU usage is constantly high.

Solution:

Use the CPU chart on Infrastructure monitoring's Hosts page to create an alert condition for system metrics.

Virtual memory capacity

Problem:

Your night shift needs to be alerted when virtual memory for a set of background workers reaches an average of 10G for at least two minutes.

Solution:

Use the Top memory consumers chart on Infrastructure monitoring's Processes page to create an alert condition for process metrics.

Limited bandwidth

Problem:

You want to monitor performance based on the average number of errors received or transmitted.

Solution:

Use the Top bandwidth chart on Infrastructure monitoring's Network page to create an alert condition for network metrics.

I/O read and write operations

Problem:

You are testing a new set of hosts in your staging environment, and you want to be notified when their read or write capacity rises above your test threshold level.

Solution:

Use the Top I/O operations chart on Infrastructure monitoring's Storage page to create an alert condition for storage metrics.

Host not reporting

Problem:

You want to be notified when we have stopped receiving data from an infrastructure agent.

Solution:

From the Hosts, Processes, Network, or Storage pages, create a host not reporting alert condition.

Processes not running as expected

Problem:

  • You want to be notified if any of the processes on your hosts stop reporting.

    OR

  • A process you expected to start on a host (such as a new program) is not actually running.

Solution:

From the Processes page (or from the Hosts, Network, or Storage pages), create a process running alert condition.

Examples: Threshold options

Use the thresholds dropdown for the selected Alert type to further define how you want to be alerted. Here are some examples of the options available.

Integrations providers

With infrastructure integrations, you can create an alert condition from your Integrations page. Depending on the type of provider selected (CloudFront, DynamoDB, EBS, etc.), options will vary from the Define thresholds dropdown; for example, bytes, errors, requests, CPU, connections, memory, records, latency, etc.

CPU, disk, load average, memory, swap

The System metrics thresholds dropdown allows you to select various criteria for CPU, disk, load average, memory, and swap metrics.

Byte size

The Network metrics thresholds provide flexibility with your business needs. Depending on the size of your network, you can easily set the threshold in bytes, KB, MB, GB, or TB.

For more help

If you need more help, check out these support and learning resources: