New Relic APM Review

It provides a single pane for operation teams to look at and get to the root cause of issues, although alerts tend to show how an entire cluster, and not just the issue-causing node, is performing.


Valuable Features

  • It's easy to install and to configure.
  • The UI interface and navigation make it easy for a novice user to quickly use the provided relevant performance application data to determine how well their application is performing and to see other areas which may require some further tuning.

Improvements to My Organization

It provides a single pane for operation teams to look at and get to the root cause of issues. This allows them to take the required corrective actions and to remove some of the blame game. People can only look at their own isolated metrics.

Room for Improvement

  • Ability to set-up maintenance windows, so silent time can be given to monitors for a defined period of time.
  • Alerts tend to show how an entire cluster is performing, and not only that a given node is having an issue. To get around this issue one can use parent-child relationships in the naming of given applications and set the alert conditions at the child levels.
  • For browser, the need to provide an interface which allows us to define a given business transaction, e.g. define all the steps of a given credit submittal. This is somewhat possible using “Funnels” with New Relic Insights.

Use of Solution

We've been using it for a little over a year and a half and have it in the following -

  • APM deployed across JBOSS, Tomcat, WebLogic, Glassfish, ActiveMQ, .Net
  • Server deployed across RedHat, CentOs, and Windows 2008/2012
  • DevOps deployment of APM and server agent using chef cookbook

Deployment Issues

APM and server agents are deployed via a chef cookbook or shell scripts. In one case, we had a very old version of Java running and had to deploy a corresponding older version of New Relic.

Stability Issues

It has been operational for over a year and a half across over 200+ applications on Java and .Net, with no negative impact to any of the systems.

Scalability Issues

Given that New Relic is a SaaS model we haven’t had any negative impact utilizing their UI to view given applications or do analytics with Insights.

Customer Service and Technical Support

Customer Service:

They are very dedicated and attentive to our needs as well as how their products can be utilized to enhance our support of given applications - 10/10.

Technical Support:

Their on-line technical support has always provided timely updates of any open tickets as well as very open to getting on a chat Webex to resolve issue. Many issues have been resolved on either the first or second exchange of notes in their ticketing system - 9/10.

Previous Solutions

We previously used HP Diagnostics and CA Introscope. We switched for several reasons -

  • UI interface required more training of the support teams to utilize the product because it wasn't as intuitive.
  • Large infrastructure to support and administer the environment.
  • Requires RUM appliance to tap the network so that we can gain insight into the end-user response.

Initial Setup

We deployed the agent via simple manual steps or via in-house written scripts.

Implementation Team

It was deployed by an in-house middleware team with vendor consulting in a few areas where the browser auto injection failed. The vendor team was very knowledgeable of their products and how to deploy and configure across an array of platforms.

Other Advice

Implement APM and server on non-production environments as well, especially an environment where you are running performance testing from. BrowserPro should be a requirement, and you should look into New Relic Insights as it is a great diagnostic tool. Make sure you keep, at a minimum, a few months worth of detailed data as well as populating it with additional custom attributes or data from other systems.









Disclosure: I am a real user, and this review is based on my own experience and opinions.
1 Comment
Lead Strategy Architect at a recruiting/HR firm with 1,001-5,000 employeesReal User

It is correct agree.
But regarding the Window period you could still disable in Application alert policies policies ---> so no alert for a number of server in this group.
Set a calendar with period/time/schedule should be better to deactivate policies/alert.

Regarding:
Alerts tend to show how an entire cluster is performing, and not only that a given node is having an issue. To get around this issue one can use parent-child relationships in the naming of given applications and set the alert conditions at the child levels.
It is correct you have to set additional name in the config file of your newrelic for your APM to "build" a parent-child relationship. It is easy to implement. Maybe the display in the dashboard should take into account this relation and show the node as sub-element of the cluster item in APM dashboard.

24 November 15
Guest
Sign Up with Email