boot_stats

AWS Lambda function that parses cloud watch logs and inserts stats into dynamodb table for further processing.

Overview

This AWS CloudFormation stack works together with erllambda_sam_boot_test stack.

boot_stats stack deploys Kinesis stream and lambda function for each code loading mode (interactive or embedded). CloudWatch logs accepted from lambda function are forwarded to the Kinesis streams which are then processed by boot_stats function to extract execution/boot time and store the stats into dynamodb table.

Repository contains some utilities to help analyse collected stats:

• dump_table.py export stats from a provided table in CSV format (see stats.csv for an example)
• execution time.ipynb little jupyter notebook to help with understanding of collected data.

Usage

Build and Deploy stack

Build lambda function as it's described in rebar3 erllambda plugin.

Run tests

Using erllambda_sam_boot_test repository generate stats by performing deploy/invoke/destroy cycle.

Export collected data

You can use dump_table.py script to export table in CSV format:

./etc/dump_table.py boot-stats-StatsTable-1C1QDZ2L4ZYMQ > ./etc/stats.csv

For further analysis you can use jupyter notebook to load and visualise exported data