In this video, we'll introduce Blackfire's Deterministic Profiling. By the end of it, you should have the big picture of Blackfire's profiler and what it can provide you with: detailed multi dimensional insights into your application's performance to help you uncover those elusive bottlenecks.
Let's start by clarifying what we mean by deterministic profiling. In short, it's the process of gathering performance data in a controlled and on demand way. Whenever you trigger a profile, Blackfire systematically measures how your code consumes resources like CPU, memory, and I/O so you can pinpoint exactly where and why your application might be slow.
Unlike Continuous Profiling, which relies on probabilistic sampling, deterministic profiling captures every function call during a transaction, giving you the most comprehensive view of your application's behavior. In a future video, we'll explore the comparisons between deterministic and continuous profiling.
As a reminder, Blackfire's deterministic stack provides three levels of performance insights. The simple monitoring traces are high level performance data samples, collected with nearly zero overhead for your application. The extended traces provide deeper insights that highlight major calls and SQL queries and cause light overhead.
The deterministic profiles are the most comprehensive observability data unit available, detailing the contribution of each function and service calls to all available dimensions , as as well as information about SQL queries, HTTP requests, cache usages, and hundreds of metrics. These offer the most extensive data, but carry the highest overhead.
They are still only triggered when you choose, so end users don't feel any impact. When you use Blackfire's profiler, you gain a multidimensional perspective on your code's runtime behavior. It measures how each function calls consumed resources, collecting large sets of observability data.
Here's a key benefit. You can use Blackfire Profiler across all of your environments, development, test, staging, and even production, without any code changes.
No need to modify your application source. No extra instrumentation to remove before going live. No performance impact on your end users, because overhead is incurred only when you trigger a profile.
That's it for today's introduction to deterministic profiling with Blackfire.
In the next few videos, we'll explore the many ways of triggering deterministic profiles, each features of Blackfire deterministic profiling in detail. We will walk you through analyzing core graphs, timeline view, recommendation and more.