Benthos | Benthos

It's boringly easy to use

Written in Go, deployed as a static binary, declarative configuration. Cloud native as utter heck.

Curl
Homebrew
Docker

# Install
curl -Lsf https://sh.benthos.dev | bash
# Make a config
benthos create nats/protobuf/aws_sqs > ./config.yaml
# Run
benthos -c ./config.yaml

# Install
brew install benthos
# Make a config
benthos create nats/protobuf/aws_sqs > ./config.yaml
# Run
benthos -c ./config.yaml

# Pull
docker pull jeffail/benthos
# Make a config
docker run --rm jeffail/benthos create nats/protobuf/aws_sqs > ./config.yaml
# Run
docker run --rm -v $(pwd)/config.yaml:/benthos.yaml jeffail/benthos

Mapping
Multiplexing
Enrichments

input:
  gcp_pubsub:
    project: foo
    subscription: bar
pipeline:
  processors:
    - bloblang: |
        root.message = this
        root.meta.link_count = this.links.length()
        root.user.age = this.user.age.number()
output:
  redis_streams:
    url: tcp://TODO:6379
    stream: baz
    max_in_flight: 20

input:
  kafka:
    addresses: [ TODO ]
    topics: [ foo, bar ]
    consumer_group: foogroup
output:
  switch:
    cases:
      - check: doc.tags.contains("AWS")
        output:
          aws_sqs:
            url: https://sqs.us-west-2.amazonaws.com/TODO/TODO
            max_in_flight: 20
      - output:
          redis_pubsub:
            url: tcp://TODO:6379
            channel: baz
            max_in_flight: 20

input:
  mqtt:
    urls: [ tcp://TODO:1883 ]
    topics: [ foo ]
pipeline:
  processors:
    - branch:
        request_map: |
          root.id = this.doc.id
          root.content = this.doc.body
        processors:
          - aws_lambda:
              function: sentiment_analysis
        result_map: root.results.sentiment = this
output:
  aws_s3:
    bucket: TODO
    path: '${! meta("partition") }/${! timestamp_unix_nano() }.tar.gz'
    batching:
      count: 100
      period: 10s
      processors:
        - archive:
            format: tar
        - compress:
            algorithm: gzip

Takes Care of the Dull Stuff

Benthos solves common data engineering tasks such as transformations, integrations, and multiplexing with declarative configuration. This allows you to easily and incrementally adapt your data pipelines as requirements change, letting you focus on the more exciting stuff.

It comes armed with a wide range of processors, a lit mapping language, stateless windowed processing capabilities and an industry leading mascot.

Well Connected

Benthos is able to glue a wide range of sources and sinks together and hook into a variety of databases, caches, HTTP APIs, lambdas and more, enabling you to seamlessly drop it into your existing infrastructure.

Working with disparate APIs and services can be a daunting task, doubly so in a streaming data context. With Benthos it's possible to break these tasks down and automatically parallelize them as a streaming workflow.

Reliable and Scalable

Benthos runs fast and processes messages using a transaction model, making it able to guarantee at-least-once delivery even in the event of crashes or unexpected server faults.

It's completely stateless, allowing for easy deployment and liberal scaling. It also exposes metrics and tracing events to targets of your choice.

At Meltwater it's enriching over 450 million documents per day with a network of more than 20 NLP services. It sounds very interesting but rest assured, it's totally drab.

Extendable

Sometimes the components that come with Benthos aren't enough. Luckily, Benthos has been designed to be easily plugged with whatever components you need.

You can either write plugins directly in Go (recommended) or you can have Benthos run your plugin as a subprocess.

It's boringly easy to use

Takes Care of the Dull Stuff

Well Connected

Reliable and Scalable

Extendable

Sponsored by the following heroes