DynamoDB vs Cosmos DB

Audits consumed read/write capacity units, detects partition throttle events, and identifies GSI backpressure limits.

# 1. Query CloudWatch for throttled requests on base table and GSIs
aws cloudwatch get-metric-data \
  --metric-data-queries '[
    {
      "Id": "m1",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/DynamoDB",
          "MetricName": "ThrottledRequests",
          "Dimensions": [{"Name": "TableName", "Value": "prod_orders"}]
        },
        "Period": 60,
        "Stat": "Sum"
      }
    }
  ]' \
  --start-time "$(date -u -v-1H +%Y-%m-%dT%H:%M:%SZ)" \
  --end-time "$(date -u +%Y-%m-%dT%H:%M:%SZ)"

# 2. Check table description for GSI capacity status and partition metrics
aws dynamodb describe-table \
  --table-name prod_orders \
  --query "Table.{Status:TableStatus,ItemCount:ItemCount,SizeBytes:TableSizeBytes,BillingMode:BillingModeSummary.BillingMode,GSIs:GlobalSecondaryIndexes[*].{IndexName:IndexName,Status:IndexStatus,RCU:ProvisionedThroughput.ReadCapacityUnits,WCU:ProvisionedThroughput.WriteCapacityUnits}}"

Measures normalized RU consumption per physical partition, detects 429 response rate spikes, and checks autoscale ceilings.

# 1. Audit normalized RU consumption percentage across partition key ranges
az monitor metrics list \
  --resource "/subscriptions/{subId}/resourceGroups/{rg}/providers/Microsoft.DocumentDB/databaseAccounts/{account}" \
  --metric "NormalizedRUConsumption" \
  --interval PT1M \
  --aggregation Maximum

# 2. Check 429 throttle rate and autoscale throughput allocation
az cosmosdb sql container throughput show \
  --account-name prod-cosmos-cluster \
  --resource-group prod-data-rg \
  --database-name core_db \
  --name orders \
  --query "{CurrentRU:resource.throughput,AutoscaleMaxRU:resource.autoscaleSettings.maxThroughput}"