By now you should have the continuous profiling up and running. But how do you actually read all that data and turn it into answers to your performance questions?
Let's take a tour of the Blackfire Continuous Profiling dashboard.
The first thing you will notice: colors.
Each color in the flame graph represents a different dimension like CPU-Time, memory, I/O, or Wall-Time. The more intense the color, the more resources that function consumes. That makes hotspots ready to spot at a glance.
Don't worry if you have trouble perceiving colors. Their shade don't matter. Only their density does.
Next up: the flame graph itself. Think of it as your application's performance fingerprint. Each rectangle is a function call.. Wider frames means higher resource usage. Taller stack mean deeper call chains.
Hover on any frame for details.
Double click to zoom in and navigate through callers and callees.
Do you prefer raw data over visuals? The table view is for you.
Here, every function is listed and sorted by resources consumption.
By default, you will see the exclusive time: the resources spent in that function alone, excluding its children.
It's perfect to quickly identify your highest resource intensive functions.
You are not stuck on a single dimension. The dashboard lets you pick any dimension you want to investigate. The available dimensions differ for each runtime
And since this is continuous profiling, you can select any timeframe, zoom in on an incident, or look at longer trends.
Finally, one of the most powerful features: comparison mode.
Pick two timeframes: before or after a deployment, or quiet hours versus peak traffic.
The continuous profiling dashboard turns raw performance data into clear insights. It helps you not just find problems, but understand them.