Skip to main content
Release Number0.6.0
Release DateJuly 23th, 2023
Release CodenameTech Preview #4
Tag0.6.0

Release 0.6.0

Main changes​

New Logo and refresh of the frontend​

Infrahub has a new logo and a new paint of colors based on a shade of blue Infrahub's logo is inspired by a tree.

Why a tree ? because trees are composed of branches like Infrahub and the leaves on a tree have some similarities to the nodes on a graph.

User Management and Authentication​

Infrahub now supports standard user management and authentication systems.

A user account can have 3 levels of permissions

  • admin
  • read-write
  • read-only

By default, Infrahub will allow anonymous access in read-only. it's possible to disable this feature via the configuration main.allow_anonymous_access or via the environment variable INFRAHUB_ALLOW_ANONYMOUS_ACCESS

Authentication mechanisms​

Infrahub supports two authentication methods

  • JWT token: Short live token that are generated on demand from the API
  • API Token: Long live token generated ahead of time.

API token can be generated via the user profile page or via the Graphql interface.

JWTTOKEN
API / GraphQLYesYes
FrontendYesNo
Python SDKSoonYes
infrahubctlSoonYes
GraphQL PlaygroundNoYes

While using the API the Authentication Token must be provided in a header named X-INFRAHUB-KEY

Groups​

With the new Group model, it's now possible to create arbitrary groups of objects. A group can is composed of members and subscribers and both can be objects of any types.

By default a CoreStandardGroup is available but it's possible to create new type of groups with additional attributes or relationships.

It's possible to retrieve the list of groups a given object is member of or subscriber of with the relationships member_of_groups and subscriber_of_groups.

These 2 relationships will be automatically added to all objects in GraphQL.

We are just scratching the surface with the groups. Over time, Groups will have a major role in infrahub and they will act as glue between a lot of other features so stay tuned for more update in the next release.

Namespace for the Schema​

All models for nodes and generics defined in the schema must now have a Namespace. The introduction of Namespace will allow us to provide a better classification of the models as the schema grows. It will also allow us to better differenciate between internal and user provided models which will ultimately lead to a better user experience.

In the process we also remove the former attribute name attribute on the model because it was highly redundant with the kind. In the new format, a model is define by its namespace and its name, which together forms for the new kind The kind doesn't have to be provided by the user anymore and its generated by concatenating the namespace and the name

Kind = Namespace + Name

In the example below, the kind of this model will be InfraRack

nodes:
  - name: Rack
    namespace: Infra
    label: Rack

Migrate your existing schema​

To migrate an existing schema, you need to:

  • define the namespace
  • delete the current name since it's not used anymore
  • rename the current kind into name
  • (optional) add a label, otherwise the label will be generated from the new kind

Impact on the GraphQL API​

Previously the name was mainly used to generate the top level graphql query. As a result of this change, the format of the GraphQL top level query are now using the Kind in CamelCase.

BeforeAfter
query {
    device {
        edges {
            node {
                id
            }
        }
    }
}
query {
    InfraDevice {
        edges {
            node {
                id
            }
        }
    }
}

Add Diff view for files in Git Repository​

It's now possible to visualise the diff for the files that are part of a Git repository in the Branch View along with the diff for the data.

Support for Memgraph as the main database​

In addition to Neo4j, it's now possible to run Infrahub on Memgraph.

The main benefits to use memgraph is its speed, especially on non-production environment.

The demo environment and the tests are now using Memgraph by default.

It's possible to start the demo environment with the flag --database=neo4j to switch to Neo4j.

Query all Generics Model directly from GraphQL​

All generics models are now exposed at the top level of the GraphQL query by default. This will make it much easier to query for generic nodes whithout having to go through the relationships of a given object.

Other Changes​

All generics objects in GraphQL

  • Python SDK : Update query methods to allow user to select fields to query and not query all relationships by default @dgarros (#748)
  • Update infrahubctl to work with authentication + various fixes @dgarros (#742)
  • Change IPHost and IPNetwork graphql type to string @wvandeun (#713)
  • Initial prototype of a synchronization engine