How to pass the image version in the notification

[radimpaseka]

Hi,
I'm trying to pass the image version in the notification when new deployment happen, I cannot find any doc what is possible. In the samples I see .app.metadata. parameter, .app.spec.source parameters, etc. is there any documentation what parameters are available?
How can I pass the parameter with image version which has been deployed?

Thx

[mubarak-j]

Have you tried fetching that value from the application status field, i.e. .app.status.summary.images?

with something like this:

template.app-sync-succeeded: |
    email:
      subject: Application {{.app.metadata.name}} has been successfully synced.
    message: |
      {{if eq .serviceType “slack”}}:white_check_mark:{{end}} Application {{.app.metadata.name}} has been successfully synced at {{.app.status.operationState.finishedAt}}.
      Sync operation details are available at: {{.context.argocdUrl}}/applications/{{.app.metadata.name}}?operation=true .
    slack:
      attachments: |
        [{
          “title”: “{{.app.metadata.name}} {{ $image := first .app.status.summary.images }}{{ $list := splitList “:” $image  }}{{ index $list 1 }}“,

[ErezWeiss]

Notice that it takes the previous image and not the new one.

[Srinivas-DKS]

Are you able to crack this to take the latest image instead of previous image

[adiii717]

This works for me

          {
            "title": "Image",
            "value": "{{index .app.status.summary.images 0}}",
            "short": true
          }

[adiii717]

Here is the configmap

apiVersion: v1
kind: ConfigMap
metadata:
  name: argocd-notifications-cm
data:
  service.slack: |
    token: $slack-token
  defaultTriggers: |
    - on-deployed
    - on-health-degraded
  trigger.on-health-degraded: |
    - description: Application has degraded
      send:
      - app-health-degraded
      when: app.status.health.status == 'Degraded'
  trigger.on-failed-operation: |
    - description: Application is Degraded.
      send:
      - app-health-degraded
      when: app.status.operationState.phase in ['Error', 'Failed']
  trigger.on-deployed: |
    - description: Application is synced and healthy. Triggered once per commit.
      oncePer: app.status.operationState.finishedAt
      send:
      - app-deployed
      when: app.status.operationState.phase in ['Succeeded'] and app.status.health.status == 'Healthy' and app.status.sync.status == 'Synced'
  template.app-deployed: |
    message: |
      {{if eq .serviceType "slack"}}:white_check_mark:{{end}} Application {{.app.metadata.name}} is now running a new version of deployment manifests.
    slack:
      attachments: |
        [{
          "title": "{{ .app.metadata.name}}",
          "title_link":"{{.app.metadata.name}}",
          "color": "#18be52",
          "fields": [
          {
            "title": "Sync Status",
            "value": "{{.app.status.sync.status}}",
            "short": true
          },
          {
            "title": "Repository",
            "value": "{{.app.spec.source.repoURL}}",
            "short": true
          },
          {
            "title": "Revision",
            "value": "{{.app.status.sync.revision}}",
            "short": true
          },
          {
            "title": "Image",
            "value": "{{index .app.status.summary.images 0}}",
            "short": true
          }
          {{range $index, $c := .app.status.conditions}}
          {{if not $index}}{{end}}
          {{if $index}},{{end}}
          {
            "title": "{{$c.type}}",
            "value": "{{$c.message}}",
            "short": true
          }
          {{end}}
          ]
        }]
  template.app-health-degraded: |
    message: |
      {{if eq .serviceType "slack"}}:exclamation:{{end}} Application {{.app.metadata.name}} has degraded.
      Application details: {{.app.metadata.name}}.
    slack:
      attachments: |
        [{
          "title": "{{ .app.metadata.name}}",
          "title_link": "{{.app.metadata.name}}",
          "color": "#f4c030",
          "fields": [
          {
            "title": "Health Status",
            "value": "{{.app.status.health.status}}",
            "short": true
          },
          {
            "title": "Repository",
            "value": "{{.app.spec.source.repoURL}}",
            "short": true
          },
          {
            "title": "Image",
            "value": "{{index .app.status.summary.images 0}}",
            "short": true
          }
          {{range $index, $c := .app.status.conditions}}
          {{if not $index}}{{end}}
          {{if $index}},{{end}}
          {
            "title": "{{$c.type}}",
            "value": "{{$c.message}}",
            "short": true
          }
          {{end}}
          ]
        }]

and this is how my app annotation look like

                  argocd-image-updater.argoproj.io/notify: "true"
                  notifications.argoproj.io/subscribe.on-deployed.slack: deployment
                  notifications.argoproj.io/subscribe.on-health-degraded.slack: deployment
                  notifications.argoproj.io/subscribe.slack: deployment

[lcc3108]

            "value": "{{index .app.status.summary.images 0}}",

In my case, the previous image was displayed.

[piodyb]

I have modified trigger.on-deployed

when: app.status.operationState.phase in ['Succeeded'] and app.status.health.status == 'Healthy' and app.status.sync.status == 'Synced' and time.Now().Sub(time.Parse(app.status.operationState.startedAt)).Minutes() >= 5

That will slow down message generation, so the .app.status.summary.images field is updated, before message template compiles. For me its working

[gecube]

@piodyb with all respect, slowing notifications down isn't a fix. It just waits until the race resolves itself.

Why summary.images shows the previous image

status.summary.images doesn't record what was deployed. It's a deduplicated, alphabetically sorted list of the images of every Pod currently in the app's resource tree, including init containers and sidecars (GetSummary, populatePodInfo).

A Deployment turns Healthy as soon as the new replicas are available. Old pods that are still terminating no longer count in status.replicas, but they are still in the tree. So when on-deployed fires, the summary contains both app:1.4.0 and app:1.5.0, and index .app.status.summary.images 0 returns whichever sorts first. That is usually the old one. The order is lexical, so 1.10.0 sorts before 1.9.0 and SHA tags land at random positions. With Istio, index 0 is simply docker.io/istio/proxyv2:... for almost every app.

A 5-minute delay works only because the old pods are gone by then. It makes every notification 5–6 minutes late (the notifications controller re-evaluates triggers on a 60s resync). It still breaks when termination takes longer, for example with a long preStop or terminationGracePeriodSeconds.

The proper way (Argo CD ≥ 3.1)

Since v3.1 the sync result stores the images of every synced resource, taken from the target manifests (argoproj/gitops-engine#642): app.status.operationState.syncResult.resources[].images. The sync operation fills this field itself, so it's already correct when on-deployed fires and no delay is needed. Tested on v3.6.1:

template.app-deployed: |
  message: |
    Application {{.app.metadata.name}} has been deployed:
    {{- range .app.status.operationState.syncResult.resources }}{{ if .images }}
    {{ .kind }}/{{ .name }}: {{ join ", " .images }}{{ end }}{{ end }}

As a one-liner for a Slack attachment field:

"value": "{{ range .app.status.operationState.syncResult.resources }}{{ if and (eq .kind "Deployment") .images }}{{ .name }}: {{ join ", " .images }} {{ end }}{{ end }}"

Note that it only covers containers (not initContainers) of resources that were part of that sync operation.

On older versions, take the version from the desired state instead of the live pods. Use .app.spec.source.kustomize.images or .app.spec.source.helm.parameters if the image is set in the Application (e.g. Image Updater with argocd write-back). If the tag lives in values in Git, use {{ (call .repo.GetAppDetails).Helm.GetParameterValueByName "image.tag" }}.

[gecube]

One more thing the 5-minute delay hides: on-deployed can fire before the rollout even starts.

A sync without hooks or sync waves completes right after apply (gitops-engine waits for resource health only in multi-step syncs). So operationState.phase becomes Succeeded while health.status is still the pre-sync Healthy, and on-deployed fires with the old pods still running. We saw this on every sync of an app without hooks: Succeeded + stale Healthy first, Progressing ~5 s later, the real Healthy ~40 s later.

After the operation the controller runs a full reconcile that writes the new health and reconciledAt in the same status update. So instead of waiting on the clock, you can wait for that reconcile:

trigger.on-deployed: |
  - when: app.status.operationState.phase in ['Succeeded'] and app.status.health.status == 'Healthy' and !time.Parse(app.status.reconciledAt).Before(time.Parse(app.status.operationState.finishedAt))
    oncePer: app.status.operationState.syncResult.revisions ?? app.status.operationState.syncResult.revision
    send: [app-deployed]

reconciledAt >= finishedAt is the same "post-sync reconcile ran" check that argocd app wait uses. The ?? in oncePer covers multi-source apps (revisions) and single-source apps (revision).

Heads-up: the notified state is keyed by a hash of when, so after changing the trigger every Healthy subscribed app gets one more notification.

[malekerramli]

ok, what if i only want the :1.0 the version part only ... not the whole url ? is there a way to truncate that string ? regex maybe ?

[malekerramli]

thanks trying this now ...

[malekerramli]

didnt work, got
failed to unmarshal template app-sync-running: error converting YAML to JSON: yaml: line 23: did not find expected key

[malekerramli]

fixed the formating issue ... it worked !

[cmanzur]

can you help me to get the image name and tag only?
I'm trying:
{{ $image := index .app.status.summary.images 0 }}{{ index (split "/" $image) 1 }}

but the notification controller shows:

level=error msg="Failed to notify recipient {slack test-argocd} defined in resource argocd/test: template: app-sync-running:2:53: executing \"app-sync-running\" at <index (split $image \"/\") 1>: error calling index: value has type int; should be string using the configuration in namespace argocd" resource=argocd/test

but it's not int, the $image is:
111111111.dkr.ecr.us-east-2.amazonaws.com/test:202f46e603271aff6ef67e5227f73e80959b3111

[cmanzur]

I found the solution:
{{ $image := index .app.status.summary.images 0 }}{{ $parts := split "/" $image }}{{ $parts._1 }}