Feature Requests

I’d like to request support for a built-in utility or plugin in Harness CI that replicates the behavior of the nrwl/nx-set-shas GitHub Action. The purpose of this utility is to accurately determine the NX_BASE and NX_HEAD SHAs required to run nx affected commands during CI pipelines. This is a common need in monorepo setups using Nx, and the existing GitHub Action cannot be used as-is in Harness due to its dependency on github.context . While we can work around this with shell scripts and GitHub API calls, having an official plugin or step in Harness would: Eliminate boilerplate scripting Avoid external GitHub Action dependencies Improve reliability and native developer experience for Nx monorepos Simplify onboarding for teams migrating from GitHub Actions Desired Behavior: When running on a feature branch, compute NX_BASE from the last successful pipeline run on When running on main, compute NX_BASE from the previous successful main run Expose NX_BASE and NX_HEAD as output variables

We need to setup a delegate configuration which is multi-region. We need to run delegates in 2 EKS clusters in two AWS regions. In each EKS cluster there will be a scalable set of delegates (say min 4 max 8) and all these delegates will use the same tag. User CI load for all users of Harness in our Harness account will use the same tag. We expect the CI load to be spread across all the delegates across the two regions. We had set this config up but the delete/cleanup tasks were not going to the same cluster where the work task was assigned to. This caused pods to not get cleaned up, resulting in resource leaks and cluster running out of resources. We have been running on a single cluster for a while and this is not an acceptable config for a critical service like CI/CD which is used across the many teams. We would like to use this multi-region ASP.

Feature request as suggest by Harness team to enable system level error retry mechanism for the below failures : - Pod eviction Failure to connect to addon client Connectivity errors between delegate and ci lite engine ref : https://support.harness.io/hc/en-us/requests/72479?page=1

Problem Statement Currently, developers must manually review static analysis reports generated in Harness pipelines. Unlike Jenkins with the Warnings NG plugin, which automatically annotates PRs with linting results, Harness lacks this inline feedback capability. This gap creates friction and delays in addressing code quality issues. Proposed Solution Introduce a native step or module in Harness CI/CD that: Parses the output of popular static analysis tools (Checkstyle, PMD, ESLint, SpotBugs, etc.). Automatically maps and posts the results as inline comments to pull requests. Optionally includes summary annotations or quality gates. Supports GitHub, GitLab, and Bitbucket SCM integrations. Is configurable for severity levels, thresholds, and comment formats. Value to Customers Streamlines code reviews by surfacing issues in-context. Increases static analysis adoption and compliance. Reduces manual effort reviewing logs or external artifacts. Matches or exceeds CI capabilities found in Jenkins (Warnings NG) and GitHub Actions (via ReviewDog). Workarounds Today While tools like ReviewDog can be used in a custom script within Harness, this adds setup complexity and lacks the seamless experience of a native integration. Ideal Outcome A reusable, native Harness feature or plugin that: Integrates seamlessly with pipeline steps. Works out-of-the-box with common linters. Posts meaningful, line-specific annotations to PRs.

you can find details on the ticket - https://support.harness.io/hc/en-us/requests/81863 we need to run our cobol build on a server, we installed docker delegate and docker harness runner. when we execute pipeline we are getting below error failed with code: 500, message: { "error_msg": "failed to create directory for host volume path: /addon: mkdir /addon: permission denied" } . we cannot use root user for any app builds on our server. We need to have delegate to run as non root user

Cloudera has requested to provide the feature unittest metrics at branch level. Currently this feature is only present for project level only.

Use Case: Right now, developers struggle to locate errors efficiently because test results are only visible within each individual pipeline. They have to manually click into each pipeline to diagnose issues, which is slowing them down. They are requesting a way to view test results in the parent pipeline when using chaining or a centralized dashboard to see all test results across nested pipelines.

You already have the chart made ( https://github.com/harness/gitness/pull/3417 ) all your missing is the index.yaml file. you wouldnt even need to host it, all you would have to do is add the index file and tar package. we could access it via raw.github.com . https://helm.sh/docs/topics/chart_repository/#the-index-file