Buildkite
Tuist Runners are currently invite-only. Reach out or ping us in the community Slack to request access for your account.
Tuist watches the self-hosted queues in your Buildkite cluster and runs the jobs it finds there on the fleet. Your pipelines keep running on Buildkite; only the compute changes. Running a job on the fleet takes three changes: connect your cluster, name a queue after a Tuist profile, and target it from a step.
-
Connect your cluster. In Buildkite, open your cluster, go to Agent Tokens, and create a token; it starts with
bkct_. In your Tuist dashboard, go to Settings → Integrations, choose Connect on the Buildkite card, and enter your organization slug and the token. The token is stored encrypted and never shown again. To change the organization or rotate the token later, edit the fields on the card and choose Save changes. Tuist begins watching the cluster within a minute. -
Name a queue after a Tuist profile. Every enabled account starts with two ready-to-use profiles,
linuxandmacos, and a job is picked up when its queue key matches a profile's label. Create self-hosted queues in your cluster keyedtuist-linuxandtuist-macos, then target one from a step:# .buildkite/pipeline.yml steps: - label: "Test" command: tuist test agents: queue: tuist-macos # or tuist-linuxyaml# .buildkite/pipeline.ymlsteps:- label: "Test"command: tuist testagents:queue: tuist-macos # or tuist-linux -
Push and watch. The job is queued, claimed by a runner, and streamed back to the Runners section of your Tuist dashboard, where you can follow its logs, steps, and machine metrics. The runner ships the log from inside the job with a credential scoped to that one job, so no Buildkite API token is needed.
Pausing and disconnecting#
Pausing dispatch on a Buildkite queue stops Tuist from taking new jobs from it. Jobs already running are unaffected.
Disconnecting the cluster on the Buildkite card under Settings → Integrations stops Tuist watching it entirely.
Cache volumes#
Keep the Xcode, module and Gradle caches and other directories between jobs with cache volumes. Add them after checkout and before the commands that use them.