If you develop or manage web applications, you'll know that maintain good technical performance It's not optional: it's the difference between satisfied users who convert… and people who abandon your site in a matter of seconds. As architectures become more complex—microservices, containers, hybrid clouds, external services—it becomes almost impossible to understand what's happening using only isolated logs or hosting graphs.
That's where New Relic comes in. This observability and APM platform lets you Monitor the behavior of your applications in real timeDetect bottlenecks, see what's happening in the infrastructure, and, increasingly important, monitor the artificial intelligence models you use in your project. All from a single dashboard without driving yourself crazy jumping between five different tools.
What is New Relic and what is it used for?
New Relic is a SaaS platform for application performance monitoring and observability (APM) Widely used globally. Its goal is to give you complete visibility into how your digital systems work: backend applications, frontend, infrastructure, databases, external services, and even artificial intelligence workloads.
To achieve this, New Relic combines in the same environment application data, infrastructure, logs, user experience, and business metricsThis allows you to move from the typical fragmented view (each thing in a different tool) to a unified approach, where it is easier to detect problems, understand their impact, and prioritize what to fix first.
The platform integrates with wide variety of modern technologiesLanguages ​​like PHP, Java, Python, and Ruby; cloud services like AWS and Azure; container environments like Kubernetes; popular databases; and frameworks of all kinds. Thanks to this compatibility, it's a very attractive option for both startups and large enterprises with complex, distributed architectures.
How New Relic works inside: the MELT model
To understand how New Relic monitors your systems, it's helpful to know the MELT model, which groups data into four main types: metrics, events, logs and tracesWith them, the platform builds a complete picture of your application's behavior.
The metrics These are numerical values ​​that describe performance: load times, CPU usage, memory consumption, throughput in requests per minute, and so on. This type of data is perfect for seeing trends over time and detecting abnormal spikes or progressive degradation.
The events These logs represent activities that occur within your applications or systems. They typically group several log lines that describe something that has happened: for example, a completed transaction, a database error, or a call to an external service. They help you understand what happened and in what context.
The logs or records These are line-by-line details of those events, with timestamps and additional attributes. Thanks to them, you can reconstruct step by step what a specific application or service has done, and they are key to investigating strange behavior or errors that are difficult to reproduce.
Finally, there are the tracesTraces show the chronological sequence of events throughout a request or entire flow. In a microservices environment, for example, a trace can traverse multiple APIs, message queues, and databases, allowing you to see where time is wasted or where errors occur within the entire chain.
The role of the New Relic agent
New Relic works thanks to a agent that is installed in your application or infrastructureThis agent instruments the code and environment to automatically collect the necessary metrics, events, logs, and traces.
Each type of monitoring typically requires its own agent: an agent for APM (applications), another for infrastructure, another for navigator (web user experience), etc. These agents are responsible for sending the data to the New Relic platform, where it is processed and displayed in visual dashboards.
Once the agent is installed and the minimum system requirements are met—for example, having administrator permission and compatible versions of operating system and programming language—, the data starts appearing in your account almost instantly, without the need for very advanced configurations.
Key metrics to monitor with APM
When you use APM tools like New Relic, there are a number of Essential metrics that should be closely monitored to understand the actual state of your applications and anticipate performance problems.
El response time This indicates how long your application takes to process a request and return a response. If this metric rises above acceptable thresholds, the user experience suffers, and you may see drops in conversion rates or service usage.
La error rate This shows how many requests end in some kind of failure, such as unhandled exceptions, HTTP 5xx responses, or errors in external services. An increase in this rate is usually a sign of problematic deployments, unstable dependencies, or internal bottlenecks.
El Use of resources (CPU, memory, disk, network) tells you how much capacity your application is consuming on the servers or containers. Excessive consumption can cause saturations, while very low usage and occasional spikes can indicate the need for optimization or scaling.
El performance in terms of throughput It is usually measured in requests per minute (RPM). This metric allows you to see how much load your application can handle, how it performs during spikes, and to what extent your infrastructure is adequately sized.
La latency It reflects the time it takes for data to travel from the client to the server and back. This is especially relevant when complex networks, CDNs, external services, or distributed architectures with multiple hops between regions or providers are involved.
La user experience It goes beyond technical numbers and can be measured with indicators like the Apdex score, which relates satisfactory and unsatisfactory response times based on a defined threshold. This type of metric helps you translate technical performance into real satisfaction for the people who use your application.
Finally, there are the business metricsNumber of transactions, revenue, conversion rate, order volume, etc. Correlating this data with technical performance allows you to clearly see how much a latency problem or a temporary drop impacts the bottom line.
Depending on the characteristics of your project, in addition to these general metrics you may be interested in monitor specific indicators, such as the timing of certain SQL queries, the performance of specific integrations, or the behavior of critical user-facing functionalities.
APM Main Dashboard: What You Can See
Once the agent is active, the APM section in New Relic gives you a powerful overview. The APM Summary page shows the most relevant metrics at a glance for each service or application.
On this page you will find, among other information, the transaction response time webs, which measures the duration of requests from the customer's point of view, although it does not always capture all the time spent on a complex operation.
Another important metric in this view is the Apdex scoreThis compares measured response times to a predefined threshold. This allows you to see what percentage of transactions are considered successful and which are not, helping you understand your users' perception of performance.
You will also see the performance in requests per minute (RPM)The error rate and host resource consumption (CPU and memory) are also considered. This combination of data makes it easier to identify correlations, for example, between traffic spikes, latency increases, and CPU saturation.
In addition to the summary, New Relic offers specific views for different components: distributed traceability, individual transactions, databases, external services or even Kubernetes environments, where you can see metrics such as the age of deployments, the status of pods and performance comparisons between different versions.
Advanced monitoring: traces, external services, and Kubernetes
The functionality of distributed traceability This view records the behavior of requests traversing multiple services or microservices. In this view, you can analyze the total duration of each trace, the spans that comprise it, the errors that occur, and the most common paths your traffic takes.
The section of Transactions This list shows the most relevant requests, ordered by percentage of real-time consumed, performance, and resource usage. By delving deeper into each transaction, you can see which parts of the code, database queries, or external calls are slowing down the application the most.
Regarding databases, New Relic shows operations, response times and throughput of the queries and statements executed. This helps you identify slow queries, poorly designed indexes, or inefficient access patterns that affect overall behavior.
The view of external services It provides performance details, response time, and error rates for upstream and downstream entities, such as third-party APIs or internal systems. This allows you to better distinguish which problems are yours and which originate from external dependencies.
If you work with KubernetesNew Relic can show you information about deployments, available pods, associated logs, and comparisons between different metrics. This makes it easier to see how configuration changes, new versions, or automatic scaling affect the actual performance of your workloads.
Alerts and Service Level Agreements (SLAs)
In addition to collecting data, New Relic lets you Define alerts and track your SLAs to ensure that the applications meet the promised service levels.
The functionality of SLA reports It gives you an overview of uptime, downtime, and performance trends over different periods (daily, weekly, monthly). This allows you to assess how the end user perceives the quality of service and whether you are meeting your contractual or internal commitments.
The alerts You are automatically notified when any metric deviates from the expected range. To do this, you define custom conditions: which data source you want to monitor (for example, a specific application, a database, or an external service), what behavior is considered anomalous (for example, a metric remaining above or below a threshold), and how the incidents should be generated.
You can also adjust the incident preferenceto decide whether you want immediate notifications for any violation or only in the case of recurring or severe problems. Regarding the warning channelsThe platform allows integration of email, Slack, OpsGenie, and other incident notification and response tools.
The alerts have an engine of Applied intelligence This helps identify potential root causes by correlating data and reducing the amount of noise you would otherwise have to analyze manually. This allows technical teams to focus on solving problems rather than filtering them out.
Dashboards and data visualization
One of the keys to New Relic is its ability to build custom control panels that group together the important indicators for your business or your technical area.
The main panel already concentrates much of the performance information onto a single screen, encoded with graphics and colors that help identify patterns at a glance. However, you can go further and design your own panels adapted to different teams, projects or geographical regions.
These dashboards allow you to adjust the design, chart size, and set of metrics displayed. For example, it is common to create a specific panel for a country or region, another for a critical production environment, and an additional one designed for business with conversion metrics correlated with technical performance.
The possibility of share panels between teams It greatly improves collaborative work: development, operations, marketing or product can look at the same data from complementary perspectives and make coordinated decisions.
Integrated transaction traces and logs
When there is a serious performance problem, having transaction traces and associated records It makes the difference between guessing and knowing exactly what's going on.
APM traces in New Relic store detailed information about your application's transactions, including database calls, segmentation by function or method, and time spent on each part of the process. This allows you to see, for example, if a request gets stuck in a code loop, a specific query, or a call to an external service.
In parallel, the platform lets you consult the logs related to each traceFrom the logs interface, you can search by keywords, attributes, or time ranges, and quickly link relevant messages to problematic transactions.
This combination of traces and logs offers a very comprehensive context for incident resolutionreducing the time it takes to find the root cause of a problem and preventing you from having to constantly jump between different tools to reconstruct what happened.
Key benefits of New Relic
Compared to other monitoring platforms, New Relic stands out for several reasons that are particularly attractive to teams of all sizes. The first is its ease of installation and configurationbecause it offers abundant documentation and a guided assistant that detects the environment, proposes the appropriate agent, and even configures it automatically in many cases.
In day-to-day use, the interface focuses on the usability and the speed of finding the root causeFrom the main panel, you can delve deeper into graphs and tables with just a few clicks, tracing the problem from an Apdex crash to the specific transaction, code segment, or database query responsible.
New Relic also adds value thanks to the client-side error loggingThis helps you understand what's happening in users' browsers: what kinds of errors occur, how often, and in what contexts. This gives you extra context to improve the user experience.
Another advantage is the detailed record of transactionsThis allows you to identify at a glance which components of your application are slower to load or consume more resources. This type of visibility is crucial for prioritizing optimizations that truly impact overall performance.
The platform's user interface itself is quite intuitive and flexibleThe visual elements, view customization, and overall design make it suitable for both highly technical users and those just starting out in monitoring.
Installing New Relic APM: Example with PHP on Ubuntu
The specific APM agent installation process varies depending on the programming language and operating system whichever you use, but the general logic remains the same: register an account, obtain the license key, install the agent, and restart the services so that it begins collecting data.
On an Ubuntu system with PHP applications (for example, when installing WordPress on a VPS), the typical procedure begins with Create an account on New Relic Once logged in, access your account settings to copy your license key. This key identifies your information on the platform.
Next you need Add the New Relic repository The system is then updated with a command that writes the corresponding entry to the package source list. The source authority is then verified by importing the GPG key provided by New Relic, and the local package list is updated.
With the repository already configured, the PHP agent package directly from the official New Relic repositories. In some cases, the installation finishes automatically; if not, and a message appears indicating that you must run the installation utility as root, simply run the recommended command to complete the process manually.
During these steps you will be asked to indicate the application name You want to monitor the service and enter the license key you copied earlier. Finally, you just need to restart the web server—for example, the corresponding Apache service—for the agent to start working.
Once this is done, it's a good idea use your app normally Wait a few minutes for the agent to generate enough data. When you return to the New Relic interface, you should start seeing charts and metrics in the APM section. For installations that aren't going to be deployed at scale, the tool itself recommends its guided installation mode, which automates many of the previous steps.
How to use New Relic APM to optimize your application
After installation, developers can Start working with APM data immediatelyNo complex configurations are required. Simply log in to the platform, go to the APM tab, and select the application you want to analyze.
A typical workflow for investigate performance problems You could start by looking at the Apdex score graph. If it falls below the acceptable level, it's a clear sign that something is wrong with the user experience.
The next step is usually to locate spikes in the web transaction timeline graphThe colors represent different components: code execution, database queries, calls to external services, and so on. Identifying which color dominates the most pronounced spikes will guide you on where to look.
Clicking on a relevant peak takes you to the Transactions pageFrom there, you can sort the entries by those with the most negative impact on Apdex. The transaction that most negatively affects user satisfaction is usually a good candidate for detailed review.
Within that transaction, you can scroll down to the section of transaction traces and open one of them to see its step-by-step breakdown. In the tracking details tab, you'll find segments highlighted in colors like red or yellow, indicating the slowest or most problematic parts.
From here on, it's about analyze whether the time spent on those segments is reasonable depending on your criteria or if they are the source of the problem. If they are, you can work on the affected code, query, infrastructure configuration, or integration to reduce the impact on response time.
Using New Relic in e-commerce environments (Adobe Commerce / Magento)
In the field of e-commerce, especially with platforms like Adobe Commerce or Magento Open SourceNew Relic has become an almost indispensable tool for keeping the online store agile and stable.
One of its most valuable uses is the identification of bottlenecksNew Relic can detect slow SQL queries, third-party extensions that degrade performance, or external API calls that slow down the checkout process. Knowing exactly which component is holding up the store allows you to take precise action instead of making blind changes.
The platform also helps to optimize loading speed Page by page, analyzing the behavior of each section of the catalog, shopping cart, or checkout. This directly impacts user experience and conversion rates, as customers tend to abandon their purchases if wait times are long.
Another critical point is the integration monitoring New Relic integrates with ERPs, payment gateways, marketing tools, and other external systems, allowing you to monitor these connections and quickly detect outages or degradations that could block orders or cause payment errors.
On key dates like Black Friday or special campaigns, New Relic is very useful for prepare for traffic peaksBy analyzing the store's behavior under load, you can size the necessary resources, adjust the cache, and minimize the risk of downtime or complete site crashes.
New Relic AIM: observability applied to artificial intelligence
With the rise of AI-based applications, New Relic has incorporated New Relic AIM, a solution designed to monitor the entire artificial intelligence stack with the same philosophy that APM applies to traditional applications.
The objective is to offer a Full end-to-end visibility in all components of an AI solution: from the services that invoke it to the models used, the intermediate data flow, the underlying infrastructure, and the associated resource consumption.
The setup is relatively quick thanks to the AI-prepared agentsThese agents do not require complex additional instrumentation. They offer built-in support for well-known models such as OpenAI or AWS Bedrock, as well as orchestration frameworks like LangChain, which greatly simplifies implementation.
With AIM you can observe AI-specific metricsThis includes metrics such as the number of requests to models, response times, token usage, user feedback, and error rates in responses. All of this is combined with traditional APM signals, infrastructure metrics, and logs, so you can see the real impact the AI ​​layer has on the rest of the application.
Integration with New Relic APM 360 It allows you to correlate AI problems with other system components. If, for example, you detect an increase in application errors and simultaneously an increase in AI failures in the unified view, you can quickly isolate the cause in the model layer and drill down to find the exact source.
This approach helps to debug AI applications faster, to optimize performance, quality and costs (for example, by controlling the use of tokens) and to ensure compliance with internal regulations and policies related to data processing and the traceability of automated decisions.
New Relic comparison with other tools on the market
Several monitoring and observability solutions are available on the market, such as Dynatrace, AppDynamics, Datadog or Elastic APMEach with its own strengths and nuances. New Relic stands out in several aspects that are interesting for many teams.
One of them is the cost-benefit ratioNew Relic offers a free version and scalable plans that make it accessible to small and medium-sized businesses, while maintaining advanced capabilities for large organizations with complex architectures.
Another point in its favor is the simplicity of implementationAlthough all tools of this type require some initial work, New Relic has placed a strong emphasis on reducing friction through guided wizards, clear documentation, and agents that work well in a wide variety of environments.
The platform is also geared towards developers as well as business profilesIt does not just display technical data, but allows you to build views and dashboards that cross-reference performance metrics with indicators of revenue, conversion, or use of functionalities, making it easier to take actionable decisions.
In the specific field of e-commerce, its most notable feature is native integration with Adobe CommerceThis simplifies the monitoring of online stores without having to set up overly complicated configurations.
Compared to other tools, Dynatrace and AppDynamics offer very powerful capabilities, but they can be more expensive and complex to manageespecially for teams with less experience in observability. Datadog, on the other hand, has a very strong focus on infrastructure and logs, while Elastic APM is very popular in open-source environments, but usually requires more advanced configuration and maintenance.
From theory to practice: a real-world case of performance improvement
To see the real impact of having a unified observability platform, the case of EveryMatrix, where its senior systems architect, Alex Bularca, had to deal with performance issues initially using up to five different tools.
By adopting New Relic as their core platform, the team began working with a single dashboard, distributed tracking, and service mapsThis combination made it possible to pinpoint a prolonged latency problem to a specific group of overloaded threads, something that was previously much more difficult to discover quickly.
The possibility of having unified data from APM to custom queries It made the resolution of complex incidents faster and more efficient, and in many cases anomalies could be detected and corrected before they even caused a visible incident in production.
Examples like this make it clear that observability is not limited to having pretty graphics, but rather that It is a direct lever for reducing downtime, improving user experience, and containing operating costs..
New Relic has thus established itself as one of the most comprehensive solutions for companies that need to monitor and optimize the performance of their applications, reduce operational risks, and offer a high-level user experience.
Its unified approach to applications, infrastructure, customer experience, and now also artificial intelligence, combined with its ease of use, makes it a very solid option for any organization that wants to have its digital platforms under control, from small websites to large-scale distributed systems. Share the guide so more users know about the topic.