AWS Fargate Management

Analyze AWS Fargate tasks, services, and resource utilization across ECS clusters.

Phase 1: Discovery

#!/bin/bash
export AWS_PAGER=""

echo "=== ECS Clusters ==="
aws ecs list-clusters --output text --query 'clusterArns[]' \
  | tr '\t' '\n' | while read ARN; do
  aws ecs describe-clusters --clusters "$ARN" \
    --query 'clusters[].[clusterName,status,registeredContainerInstancesCount,runningTasksCount,pendingTasksCount]' \
    --output text
done | column -t

echo ""
echo "=== Fargate Services ==="
for CLUSTER in $(aws ecs list-clusters --output text --query 'clusterArns[]' | tr '\t' '\n'); do
  CLUSTER_NAME=$(echo "$CLUSTER" | rev | cut -d'/' -f1 | rev)
  aws ecs list-services --cluster "$CLUSTER" --launch-type FARGATE --output text --query 'serviceArns[]' \
    | tr '\t' '\n' | while read SVC; do
    aws ecs describe-services --cluster "$CLUSTER" --services "$SVC" \
      --query "services[].[serviceName,status,desiredCount,runningCount,launchType,platformVersion]" \
      --output text
  done
done | column -t | head -20

echo ""
echo "=== Running Fargate Tasks ==="
for CLUSTER in $(aws ecs list-clusters --output text --query 'clusterArns[]' | tr '\t' '\n'); do
  CLUSTER_NAME=$(echo "$CLUSTER" | rev | cut -d'/' -f1 | rev)
  for TASK in $(aws ecs list-tasks --cluster "$CLUSTER" --launch-type FARGATE --output text --query 'taskArns[]' | tr '\t' '\n'); do
    aws ecs describe-tasks --cluster "$CLUSTER" --tasks "$TASK" \
      --query "tasks[].[taskDefinitionArn | split('/') | last, lastStatus, cpu, memory, group, connectivity]" \
      --output text &
  done
done
wait
echo "" | column -t | head -20

echo ""
echo "=== Task Definitions (Fargate-compatible) ==="
aws ecs list-task-definitions --status ACTIVE --output text --query 'taskDefinitionArns[]' \
  | tr '\t' '\n' | tail -10 | while read TD; do
  aws ecs describe-task-definition --task-definition "$TD" \
    --query 'taskDefinition.[family,revision,cpu,memory,networkMode,requiresCompatibilities[0]]' \
    --output text
done | column -t

Phase 2: Analysis

#!/bin/bash
export AWS_PAGER=""
END=$(date -u +"%Y-%m-%dT%H:%M:%S")
START=$(date -u -d "7 days ago" +"%Y-%m-%dT%H:%M:%S" 2>/dev/null || date -u -v-7d +"%Y-%m-%dT%H:%M:%S")

echo "=== CPU & Memory Utilization per Service ==="
for CLUSTER in $(aws ecs list-clusters --output text --query 'clusterArns[]' | tr '\t' '\n'); do
  CLUSTER_NAME=$(echo "$CLUSTER" | rev | cut -d'/' -f1 | rev)
  for SVC in $(aws ecs list-services --cluster "$CLUSTER" --launch-type FARGATE --output text --query 'serviceArns[]' | tr '\t' '\n'); do
    SVC_NAME=$(echo "$SVC" | rev | cut -d'/' -f1 | rev)
    {
      CPU=$(aws cloudwatch get-metric-statistics --namespace AWS/ECS --metric-name CPUUtilization \
        --dimensions Name=ClusterName,Value="$CLUSTER_NAME" Name=ServiceName,Value="$SVC_NAME" \
        --start-time "$START" --end-time "$END" --period 604800 --statistics Average Maximum \
        --output text --query 'Datapoints[0].[Average,Maximum]')
      MEM=$(aws cloudwatch get-metric-statistics --namespace AWS/ECS --metric-name MemoryUtilization \
        --dimensions Name=ClusterName,Value="$CLUSTER_NAME" Name=ServiceName,Value="$SVC_NAME" \
        --start-time "$START" --end-time "$END" --period 604800 --statistics Average Maximum \
        --output text --query 'Datapoints[0].[Average,Maximum]')
      printf "%s/%s\tCPU:%s\tMEM:%s\n" "$CLUSTER_NAME" "$SVC_NAME" "$CPU" "$MEM"
    } &
  done
done
wait

echo ""
echo "=== Auto Scaling Policies ==="
for CLUSTER in $(aws ecs list-clusters --output text --query 'clusterArns[]' | tr '\t' '\n'); do
  CLUSTER_NAME=$(echo "$CLUSTER" | rev | cut -d'/' -f1 | rev)
  for SVC in $(aws ecs list-services --cluster "$CLUSTER" --launch-type FARGATE --output text --query 'serviceArns[]' | tr '\t' '\n'); do
    SVC_NAME=$(echo "$SVC" | rev | cut -d'/' -f1 | rev)
    aws application-autoscaling describe-scaling-policies \
      --service-namespace ecs \
      --resource-id "service/${CLUSTER_NAME}/${SVC_NAME}" \
      --query 'ScalingPolicies[].[ResourceId,PolicyName,PolicyType]' \
      --output text 2>/dev/null
  done
done | column -t

echo ""
echo "=== Container Health Checks ==="
for CLUSTER in $(aws ecs list-clusters --output text --query 'clusterArns[]' | tr '\t' '\n'); do
  for TASK in $(aws ecs list-tasks --cluster "$CLUSTER" --launch-type FARGATE --output text --query 'taskArns[]' | tr '\t' '\n' | head -10); do
    aws ecs describe-tasks --cluster "$CLUSTER" --tasks "$TASK" \
      --query 'tasks[].containers[].[name,lastStatus,healthStatus]' --output text 2>/dev/null
  done
done | column -t

Output Format

AWS FARGATE ANALYSIS
=====================
Cluster/Service        Tasks  CPU-Avg  CPU-Max  Mem-Avg  Mem-Max  Scaling
─────────────────────────────────────────────────────────────────────────
prod/web-api           3      25.4%    78.2%    45.1%    62.3%    target-tracking
prod/worker            2      65.2%    92.1%    78.0%    85.4%    step-scaling
staging/web-api        1      5.1%     12.0%    20.5%    25.0%    none

Clusters: 2 | Services: 5 | Tasks: 8 running
Platform: 1.4.0 | Health: 8/8 containers HEALTHY

Safety Rules

• Read-only: Only use list-*, describe-*, and CloudWatch queries
• Never stop tasks, update services, or modify scaling without confirmation
• Parallel execution: Use background jobs for multi-service metric queries
• Costs: Large clusters with many services may incur CloudWatch API costs

Anti-Hallucination Rules

1. NEVER assume resource names — always discover via CLI/API in Phase 1 before referencing in Phase 2.
2. NEVER fabricate metric names or dimensions — verify against the service documentation or --help output.
3. NEVER mix CLI commands between service versions — confirm which version/API you are targeting.
4. ALWAYS use the discovery → verify → analyze chain — every resource referenced must have been discovered first.
5. ALWAYS handle empty results gracefully — an empty response is valid data, not an error to retry.

Counter-Rationalizations

Shortcut	Counter	Why
"I'll skip discovery and check known resources"	Always run Phase 1 discovery first	Resource names change, new resources appear — assumed names cause errors
"The user only asked for a quick check"	Follow the full discovery → analysis flow	Quick checks miss critical issues; structured analysis catches silent failures
"Default configuration is probably fine"	Audit configuration explicitly	Defaults often leave logging, security, and optimization features disabled
"Metrics aren't needed for this"	Always check relevant metrics when available	API/CLI responses show current state; metrics reveal trends and intermittent issues
"I don't have access to that"	Try the command and report the actual error	Assumed permission failures prevent useful investigation; actual errors are informative