Cloud Public API metrics

Automate scaling decisions and integrate cluster metrics into your own tooling with the Cloud Public API. Use the metrics endpoint to monitor usage and throughput programmatically and trigger scaling actions when thresholds are crossed.

Caution

Cloud Public API metrics are rough estimates, limited to a 1-hour time window, and refreshed every 15 seconds. Do not rely on them for production monitoring or alerting. For comprehensive and accurate monitoring, use the Aerospike Database metrics endpoint instead.

To get started, create a Cloud API key and bearer token by following the Cloud API usage guide.

Example: Scale clusters based on write throughput

The following script demonstrates how to poll the metrics endpoint and scale your cluster when write throughput exceeds a threshold. Adapt this pattern to your own automation requirements.

Note

This script is an example. Do not use as-is in production.

#!/bin/bash
get_latest_write_count() {
    local response=$(curl -s https://api.aerospike.com/v1/database/clusters/$ID/metrics --insecure -H "Authorization: Bearer $TOKEN")
    local latest_count=$(echo "$response" | jq -r '.transactions.writeCount.daily[-1][1]')
    echo "$latest_count"
}

# Function to get cluster details
adjust_cluster() {
    echo "Write count exceeded threshold! Fetching cluster details..."

    local response=$(curl -s https://api.aerospike.com/v1/database/clusters/$ID --insecure -H "Authorization: Bearer $TOKEN" | jq 'del(.connectionDetails.tlsCertificate)')

    local cluster_size=$(echo "$response" | jq -r '.aerospikeCloud.clusterSize')
    local replication_factor=$(echo "$response" | jq -r '.aerospikeServer.namespaces[0]["replication-factor"]')

    echo "Cluster Size: $cluster_size"
    echo "Replication Factor: $replication_factor"

    # Scale up the cluster
    scale_cluster "$cluster_size" "$replication_factor"
}

# Function to scale cluster by increasing cluster size by replication factor
scale_cluster() {
    local current_size=$1
    local replication_factor=$2
    local new_size=$((current_size + replication_factor))

    echo "Scaling cluster from $current_size to $new_size nodes..."

    # Create JSON payload for PATCH request
    local patch_payload=$(cat <<EOF
{
  "aerospikeCloud": {
    "clusterSize": $new_size
  }
}
EOF
)
    # Execute PATCH request
    local patch_response=$(curl -s -X PATCH \
        https://api.aerospike.com/v1/database/clusters/$ID \
        --insecure \
        -H "Authorization: Bearer $TOKEN" \
        -H "Content-Type: application/json" \
        -d "$patch_payload")

    # Check if the request was successful
    local http_status=$(curl -s -o /dev/null -w "%{http_code}" -X PATCH \
        https://api.aerospike.com/v1/database/clusters/$ID \
        --insecure \
        -H "Authorization: Bearer $TOKEN" \
        -H "Content-Type: application/json" \
        -d "$patch_payload")

    if [[ "$http_status" -eq 200 || "$http_status" -eq 202 ]]; then
        echo "Cluster scaling initiated successfully (HTTP $http_status)"
    else
        echo "Cluster scaling failed (HTTP $http_status)"
    fi
}

while true; do
    write_count=$(get_latest_write_count)

    if [[ "$write_count" == "null" || "$write_count" == "" ]]; then
        echo "Error: Could not retrieve write count"
    elif [[ "$write_count" -gt "$THRESHOLD" ]]; then
        echo "Write count: $write_count (THRESHOLD EXCEEDED!)"
        adjust_cluster
        echo "I only want to scale once, exiting"
        exit 0
    else
        echo "Write count: $write_count (below threshold)"
    fi

    sleep $CHECK_INTERVAL
done