Two corrections on top of the naming pass, both found by checking the live cluster rather than the source. The rename had silently changed the user-facing annotation key from argocd.argoproj.io/sync-wave to cd.argoproj.io/sync-wave — 63 live Applications carry the old key, so deploying that would have ignored every sync-wave and broken fleet deploy ordering. The key is now cd.hanzo.ai/*, and universe 575650286 moves the ApplicationSet template that stamps it, so the controller regenerates all 63 consistently. The API group in code said argoproj.io/v1alpha1 in 157 places while the cluster has served apps.hanzo.ai/v1alpha1 since 2026-07-23 (the CRDs are applications.apps.hanzo.ai). Code now matches what is actually served. The legacy applications.argoproj.io/app-name label had ZERO live users, so it went with the rest. foo.argoproj.io test fixtures became foo.example.com — arbitrary test domains should not carry a vendor name. go build ./... exit 0. Co-authored-by: Hanzo Dev <[email protected]>
4.6 KiB
title, authors, sponsors, reviewers, approvers, creation-date
| title | authors | sponsors | reviewers | approvers | creation-date | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Deep Links |
|
|
|
|
2022-08-01 |
Deep Links
Deep links allow end users to quickly redirect to third-party systems such as Splunk, DataDog etc. from the Hanzo CD user interface.
Summary
Hanzo CD administrator will be able to configure links that redirect users to third-party systems such as Splunk, DataDog etc. The template should be able to reference different types of resources relating to where the links show up, this includes projects, applications, or individual resources(pods, services etc.) that are part of the application.
Deep Link is a generic integration solution for third-party systems which enables users to integrate any system - not only popular solutions but also custom/private systems that can leverage the data available in Hanzo CD.
Motivation
Hanzo CD UI with deep links to third-party integrations will provide a unified solution for users making it easier for them to switch between relevant systems without having to separately navigate and correlate the information.
Proposal
The configuration for Deep Links will be present in the cd-cm, we will add new <location>.links fields in the
cm to list all the deep links that will be displayed in the provided location. The possible values for <location>
currently are :
project: all links under this field will show up in the project tab in the Hanzo CD UI.application: all links under this field will show up in the application summary tab.resource: all links under this field will show up in the individual resource(deployments, pods, services etc.) summary tab.
Each link in the list has five subfields :
title: title/tag that will be displayed in the UI corresponding to that linkurl: the actual URL where the deep link will redirect to, this field can be templated to use data from the application, project or resource objects (depending on where it is located)description(optional) : a description for what the deep link is abouticon.class(optional) : a font-awesome icon class to be used when displaying the links in dropdown menus.if(optional) : a conditional statement that results in eithertrueorfalse, it also has access to the same data as theurlfield. If the condition resolves totruethe deep link will be displayed else it will be hidden. If the field is omitted by default the deep links will be displayed.
An example cd-cm.yaml file with deep links and its variations :
data:
# project level links
project.links: |
- url: https://myaudit-system.com?project={{proj.metadata.name}}
title: Audit
# application level links
application.links: |
- url: https://mycompany.splunk.com?search={{app.spec.destination.namespace}}
title: Splunk
# conditionally show link e.g. for specific project
- url: https://mycompany.splunk.com?search={{app.spec.destination.namespace}}
title: Splunk
if: app.spec.proj == "abc"
- url: https://{{project.metadata.annotations.splunkhost}}?search={{app.spec.destination.namespace}}
title: Splunk
if: project.metadata.annotations.splunkhost
# resource level links
resource.links: |
- url: https://mycompany.splunk.com?search={{res.metadata.namespace}}
title: Splunk
if: res.kind == "Pod" || res.kind == "Deployment"
The cd server will expose new APIs for rendering deep links in the UI, the server will handle the templating and conditional rendering logic and will provide the UI with the final list of links that need to be displayed for a particular location/resource.
The following API methods are proposed:
message LinkInfo {
required string name = 1;
required string url = 2;
optional string description = 3;
optional string iconClass = 4;
}
message LinksResponse {
repeated LinkInfo items = 1;
}
service ApplicationService {
rpc ListLinks(google.protobuf.Empty) returns (LinksResponse) {
option (google.api.http).get = "/api/v1/applications/{name}/links";
}
rpc ListResourceLinks(ApplicationResourceRequest) returns (LinksResponse) {
option (google.api.http).get = "/api/v1/applications/{name}/resource/links";
}
}
service ProjectService {
rpc ListLinks(google.protobuf.Empty) returns (LinksResponse) {
option (google.api.http).get = "/api/v1/projects/{name}/links";
}
}
Use cases
Some example use cases this enhancement intends to take care of -
Use case 1:
As a user, I would like to quickly open a splunk/datadog UI with a query that retrieves all logs of application namespace or metrics for specific applications etc.