How do I monitor my application when scaling?

In this stage of your product's life cycle, the focus shifts towards expanding production capacity and optimizing processes to meet increasing demand.

As your application scales, you may begin to experience performance issues. As more users interact with your application, it may struggle to keep up with requests. This could see throughput (how many requests your application can process) drop as demand increases. In this instance, you will need to scale your application so that it can handle more traffic. There are two ways you can scale:

Method	Description
Vertical	Increasing your server's resources, such as RAM or CPU power. This method is recommended when your application is experiencing performance issues due to a lack of resources.
Horizontal	Adding additional servers to your application. This method is recommended when your application experiences performance issues due to heavy traffic.

Key metrics to monitor

Metric	Description	What to look out for
Throughput	How many requests your application is processing	A decrease in throughput may indicate issues such as server overload, network problems, or performance bottlenecks.
Status codes	How your application is responding to requests	An increase in error status codes (e.g., 4xx or 5xx) can indicate issues such as broken functionality, incorrect input handling, or problems with integrations.
Connections	A gauge of the connections currently being handled	A sudden increase in connection count could suggest a surge in user activity. It can also indicate your application needs more resources to handle increased traffic or improper connection handling, leading to performance issues or bottlenecks.
Error rate	Percentage of requests that result in an error	A sudden increase in the error rate may indicate underlying issues such as bugs, server problems, or inadequate error handling.

Spotting patterns

In the scaling phase, spotting patterns becomes increasingly important, as your application's increased usage makes it impossible to investigate and reproduce every anomaly.

Some key patterns to look out for:

• Increasing error rate: Investigate and debug issues to locate the error and resolve its triggers.
• Traffic is increasing: Begin horizontally scaling your application so it can keep up with requests.
• Response times are increasing: Vertically scale your application so it has more processing resources.

With AppSignal, you will have real-time access to your application's critical metrics, visualized in graphs, allowing you to spot patterns and investigate them quickly:

You can even set up triggers for key metrics to notify you when they reach a specific threshold, allowing you to take the necessary steps to maintain your application's availability.

The other stages of product development

This article is part of a series. We have a Learning Center article for each stage in the product life cycle:

• What are the stages of product development?
• How do I monitor my application during development?
• How do I monitor my application during launch?
• How do I monitor my application when scaling?
• How to maintain my application post-development

The APM that scales with you

Don't worry about surprise invoices or having to guess your usage. Thanks to our relaxed upgrade policy, AppSignal scales with you. Join us with a free trial to start monitoring your application with AppSignal. You can be set up in as little as five minutes. If you have any questions or require help, our dev-to-dev support is on hand.