Fluvio Documentation

The fluvio-artifacts-util crate provides utilities for accessing SmartModule hub packages. A Fluvio hub package uses the .ipkg extension (e.g., package-x.y.z.ipkg) and contains the following components:

• package-meta: Metadata file for the package.
• manifest.tar.gz: A compressed archive containing the SmartModule build files.
• Signature files: Files used for verifying the package integrity and authenticity.

The fluvio-compression library provides handlers for compressing and uncompressing data within the Fluvio protocol.

Workflow:

1. Producer side: Data is compressed before being sent.
2. Consumer/SPU side: Data is uncompressed using compression metadata found in the batch attributes.

Supported Compression Codecs:

• None (default)
• Gzip
• Snappy
• Zstd
• LZ4

Fluvio uses two types of Helm charts for Kubernetes deployment, which are compiled into the Cluster installer as inline resources:

• Sys Chart: Contains Custom Resource Definition (CRD) definitions. Because CRDs are cluster-scoped, you can only deploy one sys chart per Kubernetes cluster.
• App Chart: Contains definitions for Kubernetes objects like Deployments, Services, and Pods. These are name-scoped and can be deployed to multiple namespaces for multi-tenant configurations.

The Fluvio Command Line Interface (CLI) is the primary way to communicate with a Fluvio cluster. A cluster is composed of two main components that you run using the following commands:

• Streaming Controller: Run using fluvio run sc.
• Streaming Processing Unit (SPU): Run using fluvio run spu.

To simplify communication with the SC, SPU, or Kafka environments, the CLI uses Profile files to store common parameters.

To build the storage-cli from the root of the project, use the make build-storage-cli command. After building, it is recommended to set an alias to easily access the binary.

make build-storage-cli
alias scli=$(PWD)/target/release/storage-cli

To deploy the Kubernetes dashboard and a customized metric server (which disables TLS certificate checks for kind), run the following commands:

1. Apply the customized metric server configuration.
2. Apply the recommended Kubernetes dashboard deployment.
3. Apply the user configuration.

kubectl apply -f metric-server.yaml
kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.4.0/aio/deploy/recommended.yaml
kubectl apply -f user.yaml

To inspect the raw JSON configuration and state of a specific topic (e.g., a topic named test) within the default namespace, use the following kubectl command:

kubectl get --raw /apis/fluvio.infinyon.com/v1/namespaces/default/topics/test | jsonpp

By default, builds use the Rust debug profile, which is faster to build but results in larger and slower binaries. To generate optimized binaries for production environments, use the RELEASE=true flag with the Makefile.

To clean up build artifacts and generated files, use the make clean command.

$ make build-cluster RELEASE=true
$ make clean

To compile SmartModules to WebAssembly, you must add the appropriate Rust target using rustup.

• For standard compilation: use wasm32-wasip1.
• If using --nowasi flags: use wasm32-unknown-unknown.

You can retrieve and display all Infinyon objects by using kubectl get --raw against the Fluvio Infinyon API endpoint. Piping the output to jsonpp will format the raw JSON for better readability.

kubectl get --raw /apis/fluvio.infinyon.com/v1 | jsonpp