What is ECS Daemon Service Collection?
This documentation explains how to collect metrics and logs from your Amazon ECS (Elastic Container Service) infrastructure using a daemon service approach. The daemon service runs a container on each node of your ECS cluster to collect infrastructure metrics and logs, then forwards them to SigNoz for monitoring and analysis.
How It Works
The daemon service approach involves deploying a specialized container that:
- Runs as a daemon: The service deploys one container per ECS cluster node
- Collects infrastructure data: Gathers metrics and logs from the host instance and Docker containers
- Acts as a gateway: Receives OpenTelemetry Protocol (OTLP) telemetry data from your applications
- Forwards to SigNoz: Sends all collected data to your SigNoz instance for visualization and analysis
Architecture Overview
ECS Cluster Nodes
├── Your Application Containers
├── Daemon Service Container (OpenTelemetry Collector)
│ ├── Collects host metrics
│ ├── Collects container logs
│ ├── Receives OTLP data from apps
│ └── Forwards data to SigNoz
Prerequisites
Before setting up the daemon service, ensure you have:
- ECS Cluster: An active ECS cluster with at least one task definition
- Launch Type: Your ECS cluster must use either EC2 or External launch type
- SigNoz Instance: Either SigNoz Cloud account or self-hosted SigNoz installation
- AWS CLI: Configured with appropriate permissions for ECS and CloudFormation operations
- IAM Permissions: Access to create CloudFormation stacks, ECS services, and Parameter Store resources
Note for Fargate Users: If you're using Fargate launch type, this daemon service approach won't work. Instead, check out the ECS Sidecar documentation for Fargate-compatible collection methods.
What Data Gets Collected?
- Infrastructure metrics (CPU, memory, network, disk, container resource usage)
- Container and system logs (application, system, Docker daemon logs)
- Application telemetry (traces, custom metrics, OTLP-compatible data)
Launch Type
- EC2: Fully supported; collects host metrics/logs with host network access.
- External: Supported for hybrid/external resources; may need extra network setup.