An Introduction
Amazon Web Services provides a serverless computing service called Lambda that lets you run your code on a high-availability computing infrastructure. It takes care of partitioning and managing the servers for you. It can be triggered via over 200 other AWS services (Async and Event Driven), directly from your code through the Amazon API Gateway (Synchronous), or in a Poll-based system for data streams by detecting changes in the stream or state. Many companies are opting to use Lambda to run parts of their codebase as its key features are time-based pricing, meaning you only pay for the time your code is actually being run, and efficient scaling, cutting time for manually deploying more servers or partitioning existing servers.
To monitor the health and performance of your Lambda functions (as well as other AWS services you might use), AWS provides a monitoring tool called CloudWatch. However, there are some limitations. Users complain that the graphs shown sometimes don’t have the metrics needed and this requires parsing through the logs to find pertinent information. In addition, creating custom graphs requires directly editing JSON files. Another key issue brought up in reviews is that the permissions of each Lambda instance can be difficult to set up. Even having to view the permissions of a lambda instance requires access to the account hosting the Lambda instance. Additionally, users often request ways to extrapolate pricing costs for upcoming months based on expected traffic to their Lambda instance without having to calculate the costs themselves.
Lore
Our goal when creating ghost was to create an application for users of AWS Lambda to analyze metrics of their serverless compute services with features designed specifically to address common issues with AWS CloudWatch metrics and improve interaction experience with Lambda metrics. Some of our core features include:
- displaying more metrics such as throttles and concurrent executions with instant refresh
- the ability to create custom graphs with a UI
- an adjustable pricing calculator to help estimate future costs, complemented with providing past billing information
- allowing the user to access the user’s available permissions, as well as the ability to set new permissions for the specific function
Visit our website at ghost-metrics.io for more information or head over to our GitHub!
What’s next for team ghost?
Our team is actively working on adding additional testing. We currently only have integration testing for our route handlers on Mac OS, but we plan on expanding this to Windows, as well as adding unit testing. We also hope to eventually add alerts and the ability to search for a specific Lambda Function by its name or ARN (with autocomplete).
Get involved!
Ghost is an open-source product supported by the tech accelerator OS Labs. We invite you to provide feedback, contribute, and take part in developing the future of ghost. Interested in contributing to ghost or the Open Source community? More information is listed on our website. We are so excited for you to try out our application!
Meet the team
Krisette Odegard — LinkedIn | GitHub
Akash Patel — LinkedIn | GitHub