Release 2.18.00

What’s New in Release 2.18

Upgrading your installation

Our upgrade guide takes you step by step through the process of upgrading to the latest version of CircleCI Server, including any checks and updates required before starting the upgrade process.

Note: If at any time your organization name has been changed, there is a script that must be run before starting the upgrade process.

Notes and Best Practices

We now require a minimum 32GB of RAM for the Services Machine.
We have made some changes to our Redis configuration. If you have externalized Redis then you’ll need to update your configuration. Please contact your Customer Success Manager.
We have also made changes to our Postgres version and require at least postgreSQL v9.5.16. If you have externalized postgreSQL then please update to at least that version in 2.17.x before upgrading to 2.18.

Known Issues

Nomad garbage collection only runs every 4 hours. This could prevent nomad jobs from running in environments with heavy usage. If you hit this bug please follow this work around:

Create /etc/circleconfig/schedulerer/customizations on the services machine

Inside /etc/circleconfig/schedulerer/customizations enter the following:

export SCHEDULERER_NOMAD_BACKPRESSURE_MAX_PENDING_JOBS=500 
export SCHEDULERER_NOMAD_BACKPRESSURE_MAX_DEAD_JOBS=100000

Restart the picard-scheduler container:
```
sudo docker restart picard-scheduler
```

Features/Improvements

It is now possible to restrict environment variables at run time by adding security groups to contexts. Security groups are defined as GitHub teams or LDAP groups. After a security group is added to a context, only CircleCI users who are members of that security group may access or use the environment variables of the context. For more information see our guide to restricting a context.
Customers running storage drivers external to AWS will see improved routing times when searching for build Artifacts.
You can now customize the metrics that get output from CircleCI. For steps and options, see our Custom Metrics guide. Below is a short list of metrics that are included by default when enabling Custom Metrics:

Metric	Description
`circle.backend.action.upload-artifact-error`	Tracks how many times an artifact has failed to upload
`circle.build-queue.runnable.builds`	Track how many builds flowing through the system are considered runnable
`circle.dispatcher.find-containers-failed`	Track how many 1.0 builds
`circle.github.api_call`	Tracks how many api calls CircleCI is making to github
`circle.http.request`	Tracks the response codes to CircleCi requests
`circle.nomad.client_agent.*`	Tracks nomad client metrics
`circle.nomad.server_agent.*`	Tracks how many nomad servers there are
`circle.run-queue.latency`	Tracks how long it takes for a runnable build to be accepted
`circle.state.container-builder-ratio`	Keeps track of how many containers exist per builder ( 1.0 only )
`circle.state.lxc-available`	Tracks how many containers are available ( 1.0 only )
`circle.state.lxc-reserved`	Tracks how many containers are reserved/in use ( 1.0 only )
`circleci.cron-service.messaging.handle-message`	Provides timing and counts for RabbitMQ messages processed by the `cron-service`
`circleci.grpc-response`	Tracks latency over the system grpc system calls

You can now provide individual AMIs for both Remote Docker and machine executor jobs. Previously we provided the option for a single custom AMI to be used across both, but with v2.18, this expanded customization gives you greater control over versioning and dependencies to meet your individual CICD needs. See our VM Service guide for more information.

Bug fixes

Additional fixes around contexts and org renames.
Fixed an issue where occasionally volumes would fail to attach to spun up Remote docker/machine instances.
Fixed an issue where the CircleCI integration could not be installed on JIRA instances with the jira.com subdomain.
Fixed an issue where the Workflows page would still point to an old repo after renaming an organization.
Fixed issue where the Workflows UI would fail to refresh data automatically.
Improved context loading times in cases when they could cause timeouts in the UI.
Fixed an issue where contexts would cause builds to return CIRCLE_BUG.