Cycle Nomad Hosts When Disk Memory Reaches a Threshold
There should be a way to hook up a disk space metric from the nomad host to the auto scaler (or another third-party service running in the server cluster) to cause it to cycle any host that reaches a certain threshold. To prevent errors due to full nomad disks
Support scheduled pipelines on CircleCI server
Currently CircleCI server does not support scheduled pipelines. This feature request is to have this feature available on CircleCI server.
Allow Server customers to set a retention period for Build Data
Server customers should be able to set retention periods (~1-2 years) so they can ensure that their databases have predictable sizes. This will also allow migrations to run much quicker and allow customers to control how much data they want to retain in their installations.
Customize maximum build runtime to be more than 5 hours in Server
Currently, as the doc mentioned, the job maximum runtime is 5 hours, and no option to change it. https://circleci.com/docs/2.0/configuration-reference/#jobs > Jobs have a maximum runtime of 5 hours.Server customers, especially static install, they don't need to care about the price of the time of running instances. And some machine learning builds need more than 5 hours. CCI-I-1436
Allow Docker GC to skip certain images in Server
Currently, there is no way to give the docker-gc container a list of images it should not delete. Having the ability to configure this would be really helpful.
Nomad Autoscaling check - a test that runs enough jobs that would cause nomad to spin up 5-10 instances and scale down afterwards
It would be helpful as a customer to have a Nomad Autoscaling check - a test that runs enough jobs that would cause nomad to spin up 5-10 instances and scale down afterwards
Meaningful metrics in CircleCI Enterprise
At the moment the metrics are very basic and don't provide a lot if insight into how CircleCI Enterprise is behaving at the user level.We would like to have the following metrics to be able to tune our installation to our user needs and build dashboards so the developer teams can monitor their workflows.- build lifecycle metrics: succeeded, failed, queued, paused, running builds count- build duration- filter by: repo name, job name and workflow name CCI-I-917
Delete Self Hosted Runners that are inactive
The self-hosted runner page is cluttered and there's no way to remove runners that we no longer want to support. It would be nice to clean up the page to only see what I want to see
Specify the physical machine to run on for each build in Static installed Server
We have a GPU machine and some CPU machines to run the build. But only few projects use GPU resources as for building and testing machine learning app. So we want to specify the physical machine to run on for each build.The custom_resource class feature is similar but it's a different feature. CCI-I-1448
Enterprise Server - DataCenter / Azure - Support
At the moment the Enterprise Server product relies on the VM Service to dynamically spin up (or pre-prevision) VM Instances for machine and remote docker jobs. However, the VM Service only supports AWS and GCP. This leaves a gab in our offering, requiring customers to spin up X amount of runner agents to account for what the VM Service does dynamically that does not work within their own DataCenter and/or another Cloud Provider. This results in the following issues for customers: Adding additional steps that complex installation path for new and existing customers. An inability to dynamically provisioning instances based on load. Adding an additional management/maintenance steps to our already complex system