Wrecking Ball

Create synthetic load using AWS Lambda.

New deployment

1. Make sure you have valid AWS credentials.
2. pnpm install
3. pnpm run deploy

That would tell you an endpoint to produce as well as multistream function. Remember this endpoint.

Running

There are two scenarios:

1. Single stream scenario aka produce: (create a stream, and update it) in number of lambdas.
2. "Multistream": every lambda invocation updates a commit.

Produce:

To start the load process, create an HTTP request to the produce endpoint you got above.

• count tells how many simultaneous requests are issued to the Ceramic node,
• endpoint is a Ceramic node to run a client against.

curl --request POST \
  --url <<produce-endpoint>> \
  --header 'Content-Type: application/json' \
  --data '{
        "count": 100,
        "endpoint": "https://ceramic-foo.3boxlabs.com/"
}'

This would invoke 100 lambdas, each creates a stream and updates it.

Multistream:

To start the load process, create an HTTP request to the multistream endpoint you got above.

• count tells how many simultaneous requests are issued to the Ceramic node,
• endpoint is a Ceramic node to run a client against.

curl --request POST \
  --url <<multistream-endpoint>> \
  --header 'Content-Type: application/json' \
  --data '{
    "count": 200,
    "endpoint": "https://ceramic-foo.3boxlabs.com/",
    "hops": 30
}'

This would invoke multiple lambdas to create and update 200 (count) streams. Every invocation adds an update to a stream until it has 30 (hops) commits.

Scenario

src/handlers/consumer.ts contains a produce scenario to run against the Ceramic node. src/handlers/multistream.ts contains a multistream scenario to run against the Ceramic node.