All Integrations
CloudCloudWatch Metric Streams + IAM

AWS ECS Integration

Track container health, task counts, and service metrics across your ECS clusters. Full visibility for both Fargate and EC2 launch types with AI-powered deployment impact analysis.

Setup

How It Works

01

Deploy CloudFormation Stack

Launch the TigerOps CloudFormation template to create the IAM role and ECS task execution role extension. The stack deploys the TigerOps sidecar container definition to your clusters.

02

Enable Container Insights

TigerOps configures ECS Container Insights on your clusters and streams the metrics via CloudWatch Metric Streams. No changes to your existing task definitions are required for cluster-level metrics.

03

Add TigerOps Sidecar

For per-container application metrics, add the TigerOps sidecar container to your task definitions. It collects CPU, memory, and network stats from the ECS task metadata endpoint.

04

Map Services to Owners

TigerOps reads ECS service tags to map services to teams and environments. Alerts are routed to the correct Slack channel or PagerDuty schedule automatically.

Capabilities

What You Get Out of the Box

Service & Task Health

Desired vs. running task counts, task health status, and deployment progress monitoring. TigerOps alerts when running tasks fall below the desired count for longer than your threshold.

Container-Level Metrics

Per-container CPU utilisation, memory usage, and network I/O collected from the ECS task metadata endpoint — granular enough to identify a single misbehaving container in a task.

Cluster Capacity Tracking

CPU and memory reservation vs. utilisation across EC2 container instances. TigerOps predicts when your cluster will run out of schedulable capacity before tasks fail to place.

Deployment Tracking

Track ECS service deployments in real time. TigerOps overlays deployment events on metric charts so you can immediately see whether a new task definition caused a regression.

Fargate & EC2 Mode Support

Full support for both Fargate and EC2 launch types. TigerOps auto-detects the launch type per service and presents the appropriate metrics and cost context for each.

AI Deployment Impact Analysis

When a deployment completes, TigerOps AI compares pre- and post-deployment metric distributions and surfaces any statistically significant regressions with the affected metrics highlighted.

Configuration

ECS Sidecar Task Definition

Add the TigerOps sidecar to any ECS task definition for per-container metrics.

ecs-task-definition.json
{
  "family": "my-service",
  "containerDefinitions": [
    {
      "name": "app",
      "image": "my-app:latest",
      "essential": true
    },
    {
      "name": "tigerops-agent",
      "image": "public.ecr.aws/atatus/tigerops-agent:latest",
      "essential": false,
      "environment": [
        {
          "name": "TIGEROPS_API_KEY",
          "value": "${TIGEROPS_API_KEY}"
        },
        {
          "name": "ECS_CONTAINER_METADATA_URI_V4",
          "value": "${ECS_CONTAINER_METADATA_URI_V4}"
        }
      ],
      "logConfiguration": {
        "logDriver": "awslogs",
        "options": {
          "awslogs-group": "/ecs/tigerops-agent",
          "awslogs-region": "us-east-1",
          "awslogs-stream-prefix": "ecs"
        }
      },
      "cpu": 64,
      "memory": 128
    }
  ],
  "requiresCompatibilities": ["FARGATE", "EC2"],
  "networkMode": "awsvpc"
}
FAQ

Common Questions

Do I need to modify my task definitions to use TigerOps with ECS?

No, cluster-level metrics (task counts, CPU/memory reservation) are available via CloudWatch Container Insights without any task definition changes. For per-container application metrics you can optionally add the TigerOps sidecar container.

Does TigerOps support ECS on AWS Fargate?

Yes. TigerOps collects Fargate task metrics via the ECS task metadata endpoint (v4). The TigerOps sidecar is injected as an additional container in your Fargate task definition and reports CPU, memory, and network stats.

How does TigerOps handle ECS blue/green deployments?

TigerOps integrates with CodeDeploy to track blue/green deployment lifecycle hooks. It monitors the new target group health and application error rates during the bake period and can trigger an automatic rollback if a regression is detected.

Can TigerOps alert when an ECS service has fewer running tasks than desired?

Yes. TigerOps continuously compares RunningTaskCount against DesiredTaskCount and fires an alert if the gap persists for a configurable duration. The alert includes the task failure reason from ECS stopped task events.

How are ECS metrics correlated with application traces?

TigerOps correlates ECS container metrics with application traces by matching the ECS task ID to the trace resource attributes. When a container shows CPU saturation, TigerOps surfaces the slowest traces from that container in the same incident view.

Get Started

Stop Flying Blind Across Your ECS Clusters

Container health, task counts, and deployment tracking with AI anomaly detection. Set up in minutes.