The blameless blog

DevOps Monitoring Tools

Blameless
Blog home
DevOps
Emily Arnott
|

Wondering about DevOps monitoring tools? We explain what DevOps monitoring is, the tools you need, how they work, and their pros and cons.

What are DevOps monitoring tools?

DevOps monitoring tools are used to track application performance, potential system vulnerabilities, infrastructure health, and other performance metrics.

DevOps and Monitoring

DevOps refers to a methodology where development and operations teams collaborate, both contributing to a project throughout its entire lifecycle. This method is supported by a collection of best practices and cultural guidelines.

At the heart of most DevOps practices and values is making decisions based on DevOps metrics. Without understanding the details of how your project is advancing, you won’t be able to make informed decisions about how to progress. Then, when the code is in operation, you need to be able to receive information on the health of the service.

To receive all of this information from the system, you need DevOps monitoring tools. These tools collect data from your system and display it in an easy-to-parse format. That way, the various teams involved in the lifecycle can easily access information and act based on the state of the system.

Types of monitoring

There are two primary methods of monitoring systems: black box and white box. Black box monitoring refers to looking at your system from the outside, as a user would. You send requests to the service and record details of how it responds. This helps you understand how users will perceive your system and can identify issues they may have.

White box monitoring refers to getting information from within the system using telemetry. As code runs, it generates reports of how quickly and accurately it’s performing the tasks expected of it. This allows you to understand and fix the causes of issues more precisely. White box monitoring also looks at gauges of your system, like what resources are being used.

You can break these methods down further into the type of monitoring task they perform. There are four main types:

Resource monitoring: this reports on the resources your system needs to function and how much they’re being used, including RAM, CPU, and storage.

Having tools in place to monitor resources is important to prevent incidents caused by not having enough of a particular resource. For example, if your service is seeing a lot of usage and your servers don’t have enough CPU to run all of their requests, your users will get errors or very slow responses. Resource monitoring allows you to allocate more resources before this happens.

If you’re on a cloud setup, it’s likely that they’ll dynamically allocate resources to make sure you don’t have outages or slowness. However, you’ll likely incur additional costs to cover this, so keeping track of your usage is still important.

Network monitoring: this reports on the data coming in and out of your service as users make requests. It can report on the type and amount of data.

Monitoring your network is essential to understand how customers actually use your services – what sort of requests are they making, how often are they making them, and what combinations of requests are most popular. These are valuable insights for determining how impactful outages are and where to allocate development resources. As this monitoring is integrated into your DevOps setup, both development and operators can work with this information.

Network monitoring is also essential for security. You need to be able to monitor incoming traffic to prevent malicious attacks.

Application performance monitoring: this reports on how requests are being responded to. It can do this by simulating different types of requests to the system and recording how quickly and accurately they’re responded to.

This type of monitoring is important because it directly reflects the experience users are having. You can set up the monitoring tool to simulate requests and combinations of requests that are used frequently. When these monitoring tools report issues, you’ll understand exactly what problems are being encountered by users, and prioritize accordingly to keep users happy.

Third party monitoring: if you use a microservice architecture, your overall system might depend on integration with third party services. This type of monitoring checks these connections and reports on external issues that affect your system.

It’s important to continuously monitor third party components as they could experience issues that your system can’t detect directly. If a third party component is experiencing an outage, it might appear in your system as just some requests going unfulfilled. Without checking directly the status of the third party component, you won’t be able to diagnose what’s wrong with your system.

Internal process monitoring

Another type of DevOps monitoring involves tracking your internal DevOps lifecycle. Each stage of the lifecycle involves certain changes to the project, which are then checked for completion before the project moves to the next stage. For example, in the integration stage, the code must be merged into the codebase before moving on to testing.

Monitoring how quickly projects move through each stage can be greatly beneficial. If you can identify where slowdowns or roadblocks are happening, you can focus on coming up with processes to speed up these steps.

To monitor internal processes, have your system generate a log whenever code progresses from stage to stage. You can have tools that convert these logs to line graphs or other visuals that make it obvious where stoppages are happening.

Picking monitoring tools

To get a complete picture of your system, you’ll need to pick tools that provide all of these types of monitoring. Depending on the type of service you offer, different DevOps monitoring tools will be more or less useful. If you rely heavily on 3rd party components, investing more heavily in monitoring them will be necessary to make sure you aren’t caught in a chain reaction of outages.

Your monitoring tools should also be able to transform data into something meaningful and actionable. Here are some of the features to look for that may be useful:

  • Creating dashboards that can show all the most critical metrics at a glance
  • Creating graphs of how the metrics change over time, highlighting patterns and deviances
  • Triggering alerts when metrics pass some threshold
  • Logging actions and highlighting abnormal events
  • Building databases of event history that can be searched

Monitoring tools will likely be a part of a larger tool stack, so making sure your other tools can interpret the monitoring data is key.

Blameless can help interpret a variety of DevOps monitoring tools and provide you with actionable and useful data. You can transform data into SLIs and SLOs to get a direct metric of user happiness and business value. Find out how by checking out a demo!

Emily Arnott

About Emily Arnott

Emily is the Community Relations Manager at Blameless, where she fosters a place for discussing the latest in SRE. She has also presented talks at SREcon, Conf42, and Chaos Carnival. Follow Emily on Twitter.

Get the latest from Blameless

Receive news, announcements, and special offers.