Eliminates lambda cold starts
npm install serverless-plugin-flambe!npm downloads total !npm version !npm license
``sh`
npm install serverless-plugin-flambe --save-dev
orsh`
yarn add serverless-plugin-flambe --dev
In the serverless file, add serverless-plugin-flambe within the plugins entry.
Example:
`yaml
functions:
hello:
handler: src/hello.handler
timeout: 15
plugins:
- serverless-plugin-flambe
`
By default all functions will then be automatically scheduled, wrapped to accept scheduled events, and immediately invoked post-deployment. If you want more granular control, options can be configured within a custom flambe variable.
Example:
`yaml`
custom:
flambe:
regex:
- hello
- /good.*/
memorySize: 512
rate: rate(5 minutes)
The above example will keep the function hello warm as well as functions prefixed with the name good. If there are no lambdas that match in the declared list, nothing will be scheduled.
In order to fine-tune the rate, input, and flow of your code, lambdas are configured on a per-lambda basis using the field flambe
| Option | Values | Default | Description |
| :--- | :--- | :--- | :--- |
| rate | AWS rate | 5 minutes | How often the lambda is to be called |wrapper
| | String | null | The file path where a custom wrapper exists (same as a function handler definition) |input
| | Object | { flambe: true } | The event the lambda receives, when it is pinged |
#### Options Example
`yaml`
functions:
hello:
handler: handlers.hello
timeout: 10
flambe:
rate: 'rate(3 minutes)'
goodbye:
handler: handlers.goodbye
flambe:
wrapper: wrapper.logger
input:
custom: 'property'
#### Custom Wrapper(s)
If you want to build a custom wrapper instead of the default flambe wrapper, it needs to be written as a higher-order-function.
Example:
`
// wrappers.js
const logger = (original) => (evt, ctx, cb) => {
console.log('Logging event data:', JSON.stringify(evt, null, 2));
return original(evt, ctx, cb);
}
module.exports = {
logger,
};
`
Make sure serverless-plugin-flambe is placed before any plugins that compile code.
Example:
`yaml`
plugins:
- serverless-plugin-flambe
- serverless-webpack
Cost per execution: $0.0000002
Cost per memory allocated 1024MB: $0.000001667
Calls per day: 288
Calls per month: 8640
Total monthly cost per-lambda: $0.016`
_**Prices calculated using the following aws information here._
0.1.5
- Flambe log retention now uses the value set in the provider
0.1.4
- Fixed a misspelling that was preventing options stage from being resolved
0.1.3
- Default to options stage, region instead of provider
0.1.2
- Re-enabled cleanup phase post local invocation
0.1.1
- PATCH: Using relative paths for original handlers instead of copying the source
0.1.0
- Changed how flambe custom options are defined
- Decrease memorySize of generated function to 128 MB
- Using Serverless lambda invocation instead of custom