title:"How to: Upgrade Deprecated Docker Images in Classic Pipelines"

description:"Find and upgrade deprecated Docker v1 and schema 1 images to support docker-27 runner in Classic Pipelines."

group:kb

sub-group:articles

toc:true

kb:false

ht:true

common:false

categories:[Pipelines]

support-reviewed:2025-07-14 AA

As part of our ongoing platform upgrades, we’re moving to a newer classic runner version based

on docker ver 27. This change deprecates older Docker image formats, and you may need to

identify and upgrade deprecated images in your pipelines.

This guide will help you:

1. Identify deprecated Docker images in your Classic Pipelines. (Step 1)

2. Upgrade those images to a modern, supported format. (Step 2)

There are two recommended approaches to identify deprecated images in your Classic

Pipelines.

This method scans historical build logs for deprecation warnings.

**How it works:**

Latest versions of dind print deprecation warning into build logs each time Docker daemon

tries to pull a deprecated image.

**Example:**

Use the following script to scan logs and extract deprecated image names:[Find deprecated images used in previous builds](https://gist.github.com/francisco-cocozza/6046028184cc12b5ee4513bdcb4217c5)

💡 For large-scale environments with thousands of builds, we offer a more scalable,

metrics-based detection method. See Appendix I for details.

Once you’ve identified deprecated images using engine metrics, the next step is to upgrade

those images to a modern format (OCI or Docker manifest v2, schema 2). This is required to

ensure compatibility with docker-27 and future releases.

Re-pushing an image using a modern Docker client will automatically convert its manifest to a

supported version.

You can automate this process with a Codefresh pipeline like the one below:

This pulls the deprecated image and re-pushes it under a new name/tag with a compliant

manifest format.

💡 You can repeat this process for each deprecated image discovered via metrics or logs.

This method enables scalable detection of deprecated images by using real-time engine

metrics. Recommended for high-volume or automated environments.

Starting from Runtime v7.5.0, engine emits

codefresh_engine_deprecated_images_pulled_total{ account_name,

pipeline_id, workflow, image_name } metric, which should be monitored to detect

deprecated images pulled within DIND.

This metric is a counter increased by 1 each time Docker daemon pulls a deprecated image.

Set the following values for the Hybrid Runtime Chart:

Before normal shutdown, the engine pod waits for the scrape of the metrics. This helps ensure

that the latest metric values are not missed.

The timeout is controlled by the METRICS_PROMETHEUS_SCRAPE_TIMEOUT engine variable. If

you want to be sure that the Prometheus operator will scrape the latest metric values before the

engine exits, we recommend setting this variable to 4 times your scrape interval.

To do so, adjust your Hybrid Runtime Chart values:

Installation of the monitoring stack is up to the customer and not covered by our support.

This section provides very basic instructions suitable for the purpose of this doc, however, they

may not meet customer’s long-term production needs and may require some adjustment by the

end user.

If the monitoring stack is not installed in the target cluster, the easiest way is to install

kube-prometheus-stack Chart:

[kube-prometheus-stack 75.6.2 · prometheus/prometheus-community](https://artifacthub.io/packages/helm/prometheus-community/kube-prometheus-stack)

While installing, please consider enabling persistence for the Prometheus operator so that

collected metrics are not purged on node rotation. Please follow the[official docs](https://artifacthub.io/packages/helm/prometheus-community/kube-prometheus-stack?modal=values&path=prometheus.prometheusSpec.storageSpec) for details

(check prometheus.prometheusSpec.storageSpec value).

**kube-prom-storage-class.yaml**

**kube-prom-values.yaml**

Please follow instructions provided by Helm after successful installation to get access to the

newly installed Grafana instance:

You can now log in to Grafana athttp://localhost:3000 as an admin user. Double-check the

instructions above for the password.

To ensure that engine metrics are collected correctly, run some build on the Runtime

under monitoring. Login to Grafana, open “Explore → Metrics” in the left-hand menu.

Type codefresh_ in “Search” field, ensure that time range covers at least one running

build. You will see a list of available metrics.

{% include

image.html

lightbox="true"

file="/images/troubleshooting/before-verify.png"

url="/images/troubleshooting/before-verify.png"

alt="codefresh_no_such_file_directory.png"

max-width="70%"

%}

codefresh_engine_deprecated_images_pulled_total metric is a counter increased

by 1 each time Docker daemon pulls a deprecated image. Labels {account_name,

pipeline_id, workflow, image_name} allow you to group results.

Easiest way to analyze data — to use the attached Grafana dashboard.

1\. Download[this Grafana dashboard json](../../grafana-dashboard.json) <br>

2\. Login to Grafana <br>

3\. Open “Dashboards → New → Import” <br>

{% include

image.html

lightbox="true"

file="/images/troubleshooting/import-dashboard.png"

url="/images/troubleshooting/import-dashboard.png"

alt="codefresh_no_such_file_directory.png"

max-width="70%"

%}

4\. Upload Dashboard JSON <br>

5\. Select “Select a JSON API data source → Mimir API”. Click “Import”

{% include

image.html

lightbox="true"

file="/images/troubleshooting/select-api-source.png"

url="/images/troubleshooting/select-api-source.png"

alt="codefresh_no_such_file_directory.png"

max-width="70%"

%}

6\. Done! 🎉

Deprecated Images Dashboard overview

{% include

image.html

lightbox="true"

file="/images/troubleshooting/deprecated-images-overview.png"

url="/images/troubleshooting/deprecated-images-overview.png"

alt="codefresh_no_such_file_directory.png"

max-width="70%"

%}

On the screenshot above:

1. Filter by account name.

2. Time range.

3. Account name (build owner). Link leads to the Admin Panel.

4. Deprecated image name.

5. Pipeline ID where pull was detected.

You can query codefresh_engine_deprecated_images_pulled_total metric directly.

Consider the following query, that returns per second rate of deprecated images pulls,

aggregated by account ID, pipeline ID, build and image:

{% include

image.html

lightbox="true"

file="/images/troubleshooting/deprecated-images.png"

url="/images/troubleshooting/deprecated-images.png"

alt="codefresh_no_such_file_directory.png"

max-width="70%"

%}

Deprecated images docker.io/tutum/dnsutils:latest and docker.io/docker/whalesay:latest

has been pulled in Pipeline #67867cfe8307bd8f9b7b034e