We're excited to announce the release of Metrics Aggregation, a new feature that allows you to aggregate your profiling metrics across processes by grouping them by specific labels. Imagine it as the sum by operator in Prometheus, but for all your profiling metrics!
Consider the following example: You have an API service that is running on multiple pods with label set service=api, and you want to know the total CPU usage across all those pods. Previously, you would have to filter for the service=api label set and then sum the CPU usage across all the pods manually.
With the new Metrics Aggregation, you can now filter for the service=api label set and sum the metrics by the service label to get the total CPU usage across all the pods that have the service=api label set.
If you're someone who runs a large infrastructure with hundreds of processes, then this feature is for you! It allows you to quickly get an overview of the resource usage across all your processes by grouping them by specific labels.
Here is an example from our demo environment, it went from
to
Which is a lot more usable compared to the previous one where it is more like finding a needle in a haystack!
How to use it
1. While in the "Profiler" tab in the Polar Signals dashboard, select the profile type and the query filters like you usually do.
2. In the "Sum by" input, select the label(s) which you want to aggregate the metrics by:
3. Click the "Search" button and that should load aggregated metrics in the graph below.
We hope this will help you make a better use of your profiling metrics and get better insights out of it. We are always eager to simplify your continuous profiling journey with Polar Signals Cloud (14-day free trial) and if you have any feedback about this feature, we would be happy to hear about it at our Discord community.