Moving to elastic cache serverless

Question

Currently we are using AWS elastic cache in one of our production account. We are thinking to moving to elastic cache serverless, believing it will reduce a part of our bill. 
While using aws calculator for determine the cost for elastic cache serverless we need some data from our current work load in the production account.
1.Average cache data size (in GB)
2.Average simple request rate
3.Average data transferred per request [KBs]

From the production account i have only the metric data related to AWS elastic cache, How can i get or calculate the above 3 data for the current workload i have in my production account. So that i can calculate the cost for AWS elastic cache serverless and ensure it will save cost or not

Oleksii Bebych · Accepted Answer

Example Script to Gather Metrics

```
import boto3
from datetime import datetime, timedelta

# Initialize boto3 clients
cloudwatch = boto3.client('cloudwatch')

# Define the time period
end_time = datetime.utcnow()
start_time = end_time - timedelta(days=7)  # last 7 days

def get_metric_statistics(metric_name, namespace, period, start_time, end_time, statistics, dimensions):
    response = cloudwatch.get_metric_statistics(
        Namespace=namespace,
        MetricName=metric_name,
        StartTime=start_time,
        EndTime=end_time,
        Period=period,
        Statistics=statistics,
        Dimensions=dimensions
    )
    return response['Datapoints']

# Get average cache size in GB
bytes_used_for_cache = get_metric_statistics(
    metric_name='BytesUsedForCache',
    namespace='AWS/ElastiCache',
    period=3600,
    start_time=start_time,
    end_time=end_time,
    statistics=['Average'],
    dimensions=[{'Name': 'CacheClusterId', 'Value': 'your-cluster-id'}]
)
average_cache_size_gb = sum([datapoint['Average'] for datapoint in bytes_used_for_cache]) / len(bytes_used_for_cache) / (1024 ** 3)

# Get average request rate
cmd_get = get_metric_statistics(
    metric_name='CmdGet',
    namespace='AWS/ElastiCache',
    period=3600,
    start_time=start_time,
    end_time=end_time,
    statistics=['Sum'],
    dimensions=[{'Name': 'CacheClusterId', 'Value': 'your-cluster-id'}]
)
cmd_set = get_metric_statistics(
    metric_name='CmdSet',
    namespace='AWS/ElastiCache',
    period=3600,
    start_time=start_time,
    end_time=end_time,
    statistics=['Sum'],
    dimensions=[{'Name': 'CacheClusterId', 'Value': 'your-cluster-id'}]
)
average_request_rate = (sum([datapoint['Sum'] for datapoint in cmd_get]) + sum([datapoint['Sum'] for datapoint in cmd_set])) / (7 * 24 * 3600)

# Get average data transferred per request in KB
network_bytes_in = get_metric_statistics(
    metric_name='NetworkBytesIn',
    namespace='AWS/ElastiCache',
    period=3600,
    start_time=start_time,
    end_time=end_time,
    statistics=['Sum'],
    dimensions=[{'Name': 'CacheClusterId', 'Value': 'your-cluster-id'}]
)
network_bytes_out = get_metric_statistics(
    metric_name='NetworkBytesOut',
    namespace='AWS/ElastiCache',
    period=3600,
    start_time=start_time,
    end_time=end_time,
    statistics=['Sum'],
    dimensions=[{'Name': 'CacheClusterId', 'Value': 'your-cluster-id'}]
)
total_network_bytes = sum([datapoint['Sum'] for datapoint in network_bytes_in]) + sum([datapoint['Sum'] for datapoint in network_bytes_out])
total_requests = sum([datapoint['Sum'] for datapoint in cmd_get]) + sum([datapoint['Sum'] for datapoint in cmd_set])
average_data_per_request_kb = (total_network_bytes / total_requests) / 1024

print(f"Average Cache Size (GB): {average_cache_size_gb}")
print(f"Average Request Rate (requests/second): {average_request_rate}")
print(f"Average Data Transferred per Request (KB): {average_data_per_request_kb}")
```

Moving to elastic cache serverless

Relevant content