Deep Links
Deep links allow users to quickly redirect to third-party systems, such as Splunk, Datadog, etc. from the Argo CD user interface.
Argo CD administrator will be able to configure links to third-party systems by providing deep link templates configured in argocd-cm
. The templates can be conditionally rendered and are able to reference different types of resources relating to where the links show up, this includes projects, applications, or individual resources (pods, services, etc.).
Configuring Deep Links
The configuration for Deep Links is present in argocd-cm
as <location>.links
fields where <location>
determines where it will be displayed. The possible values for <location>
are:
project
: all links under this field will show up in the project tab in the Argo CD UIapplication
: all links under this field will show up in the application summary tabresource
: all links under this field will show up in the 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 corresponding application, project or resource objects (depending on where it is located). This uses text/template pkg for templatingdescription
(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 menusif
(optional): a conditional statement that results in eithertrue
orfalse
, it also has access to the same data as theurl
field. If the condition resolves totrue
the deep link will be displayed - else it will be hidden. If the field is omitted, by default the deep links will be displayed. This uses antonmedv/expr for evaluating conditions
Note
For resources of kind Secret the data fields are redacted but other fields are accessible for templating the deep links.
Warning
Make sure to validate the url templates and inputs to prevent data leaks or possible generation of any malicious links.
As mentioned earlier the links and conditions can be templated to use data from the resource, each category of links can access different types of data linked to that resource. Overall we have these 4 resources available for templating in the system:
app
orapplication
: this key is used to access the application resource data.resource
: this key is used to access values for the actual k8s resource.cluster
: this key is used to access the related destination cluster data like name, server, namespaces etc.project
: this key is used to access the project resource data.
The above resources are accessible in particular link categories, here’s a list of resources available in each category:
resource.links
:resource
,application
,cluster
andproject
application.links
:app
/application
andcluster
project.links
:project
An example argocd-cm.yaml
file with deep links and their variations :
# sample project level links
project.links: |
- url: https://myaudit-system.com?project={{.project.metadata.name}}
title: Audit
description: system audit logs
icon.class: "fa-book"
# sample application level links
application.links: |
# pkg.go.dev/text/template is used for evaluating url templates
- url: https://mycompany.splunk.com?search={{.app.spec.destination.namespace}}&env={{.project.metadata.labels.env}}
title: Splunk
# conditionally show link e.g. for specific project
# github.com/antonmedv/expr is used for evaluation of conditions
- url: https://mycompany.splunk.com?search={{.app.spec.destination.namespace}}
title: Splunk
if: application.spec.project == "default"
- url: https://{{.app.metadata.annotations.splunkhost}}?search={{.app.spec.destination.namespace}}
title: Splunk
if: app.metadata.annotations.splunkhost != ""
# sample resource level links
resource.links: |
- url: https://mycompany.splunk.com?search={{.resource.metadata.name}}&env={{.project.metadata.labels.env}}
title: Splunk
if: resource.kind == "Pod" || resource.kind == "Deployment"