Set Up Performance

With performance monitoring, Sentry tracks your software performance, measuring metrics like throughput and latency, and displaying the impact of errors across multiple systems. Sentry captures distributed traces consisting of transactions and spans, which measure individual services and individual operations within those services. Learn more about our model in Distributed Tracing.

Configure the Sample Rate

Sampling for transactions must also be configured before

tracingThe process of logging the events that took place during a request, often across multiple services.
is enabled in your app. Set the sample rate for your transactions by either:

  1. Setting a uniform sample rate for all transactions using the traces_sample_rate option in your SDK config to a number between 0 and 1. (For example, to send 20% of transactions, set traces_sample_rate to 0.2.)
  2. Controlling the sample rate based on the transaction itself and the context in which it's captured, by providing a function to the traces_sampler config option.

The two options are meant to be mutually exclusive. If you set both, traces_sampler will take precedence.

config/packages/sentry.yaml
Copied
sentry:
    options:
        # Specify a fixed sample rate:
        traces_sample_rate: 1.0
        # Or provide a custom sampler:
        traces_sampler: 'sentry.callback.traces_sampler'

# Only needed when using the `traces_sampler`
services:
    sentry.callback.traces_sampler:
        class: 'App\Service\Sentry'
        factory: [ '@App\Service\Sentry', 'getTracesSampler' ]

The service needed for the traces_sampler option can be implemented as follows:

src/Service/Sentry.php
Copied
<?php

namespace App\Service;

class Sentry
{
    public function getTracesSampler(): callable
    {
        return function(\Sentry\Tracing\SamplingContext $context): float {
            // return a number between 0 and 1
        };
    }
}

Learn more about how the options work in Sampling Transactions.

Verify

While you're testing, set traces_sample_rate to 1.0, as that ensures that every transaction will be sent to Sentry.

Once testing is complete, we recommend lowering this value in production by either lowering your traces_sample_rate value, or switching to using traces_sampler to dynamically sample and filter your transactions.

Connecting Services

If you are also using Performance Monitoring for JavaScript, depending on where your request originates, you can connect traces:

  1. For requests that start in your backend, by adding a meta tag in your HTML template that contains
    tracingThe process of logging the events that took place during a request, often across multiple services.
    information.
  2. For requests that start in JavaScript, by the SDK setting a header on requests to your backend.

Otherwise, backend services with Performance Monitoring connect automatically.

Improve Response Time

Response time is somewhat impacted when you use the performance capabilities in your PHP application, (depending on the traces_sample_rate you've configured). This is due to the nature of PHP, where requests are usually sent as part of the process in which you serve your users' response. This affects the time it takes to send a request to Sentry because it's added to your servers' response time.

To offset this and make the addition close to zero, run Relay locally on the same machine or a local network that can act as a proxy/agent. Doing this will make the PHP process send requests to your local Relay, which will then forward them to Sentry, instead of sending requests to Sentry directly.

To begin using Relay, check out our docs for getting started. We recommend using Relay in managed mode (which is the default).

Follow the instructions in the Relay docs to send a test event through Relay to Sentry. Don't forget to update your DSN to point to your running Relay instance. After you set up Relay, you should see a dramatic improvement to the impact on your server.

Next Steps:

Help improve this content
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) to suggesting an update ("yeah, this would be better").