Building a robust monitoring system part 1

Building a robust monitoring system part 1

What is a “robust” monitoring system

A “robust” monitoring system is a system which not only keeps track of the server health but the behavior of the application and the customers also. So what does it take to complete a monitoring system? These are mentioned below:

  1. System health monitoring like RAM, CPU, Load average etc.
  2. Application behavior i.e throughput and latency
  3. The behavior of client device namely Browser/mobile app
  4. To check the health of API you can integrate an automated suite to check the functionality of the API
  5. Application profiling to check which transaction is causing the exact issue
  6. Plugins to monitor the respective aspect of various servers e.g Innodb buffer in case of MySql and number of messages in the queue in rabbitmq
  7. Ability to detect slow calls and slow queries

How to build a monitoring system

This will be a multi-part tutorial, each covering one key monitoring tool and best practices for the same.

Monitor server resources with New Relic

We have already covered Zabbix installation to monitor the system resource and let you know any incident of resource depletion. Zabbix is open source and entirely free but there are some issues

  1. Management of Zabbix server
  2. Agent installation
  3. Addition of the client to the server
  4. Patching your server with the latest stable version to get more features
  5. The most hectic task is monitoring the monitoring system. You need to check the status of Zabbix also else you may miss out critical alerts

If you want to get rid of the above-mentioned issue you can opt for paid tools like new relic and appdynamics. The beauty of these tools are given below:

  1. Precompiled agents with good support and documentation
  2. Very good UI console with nice and clear dashboard
  3. Industry standards metrics are readily available
  4. Easy alert creation
  5. Ability to integrate several sources to create a correlation dashboard

To install the new relic infra client please follow the steps mentioned in New Relic infra installation . Once done with the installation go to infra section in New Relic you will your host is added there and system stats data is flowing in.

New relic infrastucture

As you can see you have all the system data readily available to monitor the system and set alerts as per your required thresholds.

Creating alarms in New Relic

Alarm creation in New Relic quite straightforward.

  • Select the metric on which you want the alert

New Relic alert

  • Give condition a name like “high-cpu-usage”
  • Select the hosts
  • Set the thresholds

new relic alert condition

  • Name the policy like new-relic-infra-policy

New relic infrastructure policy

  • Add the desired email id in the recipient.

Once done with mentioned steps you are done with the new relic infra installation and achieved a milestone in monitoring your system.

In our next blog, we will set up New Relic APM to monitor the behavior of our application.

One Reply to “Building a robust monitoring system part 1”

Leave a Reply

Your email address will not be published. Required fields are marked *