mathpaquette / tskmgr Goto Github PK

Affinity feature was created to execute tasks on the same runners as they were executed within previous builds. This feature brings a lot of complexity to the code and prevents run to use all available runners scheduling most of the tasks to one of them. Even regardless of the described problem, it appears that this feature does not provide any significant functionality or impact on the application's performance or behavior, so this feature could be safely decommissioned and removed from tskmgr.

0 = green
1 = red

Add feature that give us possibility to execute dependency tasks first, and make dependent tasks wait for successful completion. In case of failure all dependent tasks should be aborted.
Example:
Project1 depends on Project2. Project2 depends on Project3.

• Tskmgr should execute Project3's task, then after successful completion execute Project2's task and only then Project3's task.
• If Project3's task failed then Project2's and Project1's task should be aborted.

Is this meant to be a drop in replacement for nx-cloud? If so, do u have an example project?

Add possibility to authenticate on PostgreSQL server using mTLS.

Ag-Grid on Task Details page is losing the scroll position after the grid data update.

I would like to store task log files in NFS shared folder that requires Kerberos authentication. For that I need to mount this
folder as a volume to the API image. Mounting NFS shares with Kerberos supported by docker, but Kerberos part requires Kerberos libraries installed inside the docker image.
Task:

• Add Kerberos libraries to API docker image.

in execution, tasks are getting overlapped

On the Runs page, in the "Last Update" column the maximum value is ">1d ago" which is not enough to distinguish runs which were done a few days ago. The behaviour of the column should be changed in a way that shows the number of days for runs executed less than 31 days ago, and >30d ago otherwise.

Docker images could not been build

• pull requests per day
• runs per day (failed,completed,aborted, run type)
• most frequent failed tasks
• efficiency (Total time / # runners / run time = efficiency %)

To improve the health of the code in tskmgr monrepo eslint-plugin-deprecation should be installed and enabled to through linter errors by default.
Task:

• Install eslint-plugin-deprecation.
• Configure eslint to throw an error for any detested deprecated code.
• Fix existing deprecated issues.

While running tskmgr client with node I am getting ECONNREFUSED error message.
To fix an issue node-fetch should be updated to latest version.
Related topic on Stackoverflow https://stackoverflow.com/questions/72390154/econnrefused-when-making-a-request-to-localhost-using-fetch-in-node-js

Upgrade monorepo packages

Plain JavaScript could replace lodash functionality. Regarding this, and the fact that we use lodash only in a few places in monorepo, replacing the usage with javascript functionality is better.

when failFast enable, when one task marked has failed, we need to automatically abort running tasks,.