New Relic checks
The New Relic module turns the golden signals New Relic already measures into Excubius checks: Apdex, error rate, response time and throughput of your APM services, and the success rate and failing locations of your synthetic monitors. Because they are catalog checks, their thresholds follow your entity tree and their alerts go through the same inbox and channels as everything else.
What it monitors
| Area | Checks | Why it matters |
|---|---|---|
| APM services | Apdex (warning below 0.85, critical below 0.7); error rate (warning above 2%, critical above 5%); average response time (warning above 1,000 ms, critical above 3,000 ms); throughput against its own history | Users feel a slow or failing application before any server alarm goes off. A sudden drop in throughput usually means traffic no longer reaches the application |
| Synthetic monitors | Failing locations (one is a warning, two or more critical); success rate (warning below 98%, critical below 90%) | One failing location can be the probe's network; two or more mean the endpoint is failing |
Host metrics from New Relic, CPU, memory and disk, do not have checks of their own: they feed the Windows and Linux host checks. New Relic's own issues arrive as alerts with a link back to New Relic.
Every check, with its rule, cause and action, is in the reference for New Relic checks.
Before you begin
- The New Relic source, with a user key of a read-only user and APM services, Synthetic monitors and Golden metrics selected. See Connect New Relic.
- The source kind is on. The feature
sources.newreliccovers the source and this module.
Add a resource
You do not add these resources by hand. The New Relic source adds APM services as Application (APM) resources and monitors as Synthetic monitor resources under its entity. Their checks appear the first time the source reports their metrics.
Tune a threshold for one application, for example its response time, as an override on that resource or on its entity. See Thresholds and baselines.