Agent Deployment Guide
Understanding the Polar Signals Architecture
Polar Signals provides a cloud platform for continuous profiling, helping teams gain deep visibility into their applications' performance. At the core of this system is the Polar Signals Agent, which must be deployed as a DaemonSet on various hosts to collect profiling data efficiently. The agent communicates with the server via gRPC, ensuring high-performance, low-latency data transmission for seamless profiling insights.
Continuous Profiling takes a different approach compared to traditional profiling. Instead of capturing short bursts of high-resolution data, it continuously profiles every process in your infrastructure at a relatively low sampling frequency (19Hz). This method ensures that enough data is gathered over time to build statistical significance while maintaining a low overhead of less than 1% with the Polar Signals Agent.
Deployment Options
Below is a table outlining the available deployment options for the Polar Signals Agent across different host types:
| Host Type | Description | Setup Guide |
|---|---|---|
| Kubernetes | Deploy the agent as a DaemonSet to profile workloads running on a Kubernetes cluster. | Kubernetes Setup Guide |
| Docker | Run the agent as a standalone container to profile applications within a Docker environment. | Docker Setup Guide |
| Snaps | Use Snapcraft packaging to deploy the agent on systems that support Snaps, such as Ubuntu-based distributions. | Snaps Setup Guide |
| ECS | Install the agent on Amazon ECS (Elastic Container Service) to monitor containerized workloads efficiently. | ECS Setup Guide |
For detailed instructions on deploying the Polar Signals Agent, follow the links in the table above.