Skip to content

agent-job-0.6.2

Choose a tag to compare

View all tags
@github-actions github-actions released this 13 Jul 11:37
agent-job-0.6.2
ff93e39
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Updated

  • Bumping chart version to v0.6.2 for scalr-agent v1.2.0

Added

  • Added a node-local disk backend for the cache directory (persistence.cache.hostPath.{path,type}), used when persistence.cache.enabled is true and hostPath.path is set (empty by default). The cache volume in task pods is then a hostPath mount that takes precedence over persistentVolumeClaim, while the controller keeps an ephemeral (emptyDir) cache volume. See hostPath cache guide for details.
  • Added agent.binaryCache.sizeLimit (default 5Gi) and agent.binaryCache.thresholdDays (default 10) to configure garbage collection of the local binary cache (OpenTofu, Terraform, Terragrunt, OPA, Checkov). Binaries unused for more than thresholdDays are removed, and least-recently-used binaries are evicted once the cache exceeds sizeLimit. The size limit is soft: a binary in use by an active run is never evicted.

Changed

  • The in-container bind mount for the binary cache moved from /opt/bin to /opt/binary-cache.

  • The controller and worker containers now use the image's built-in ENTRYPOINT/CMD instead of the previously hard-coded command: ["python", "-m", "tacoagent.cmd"] in the chart. Optional overrides are available via agent.controller.command, agent.controller.args, task.worker.command, and task.worker.args, but default to empty and using the image default is recommended — overriding bypasses the image's tuned startup and may break the container or degrade its performance.

Note: incorrectly built custom images with a modified or missing default entrypoint may be affected. If you package custom scalr/agent images, ensure the default entrypoint is not stripped or overridden.

Fixed

  • Fixed helm template/helm install failing with YAML parse error ... block sequence entries are not allowed in this context when persistence.data.persistentVolumeClaim.storageClassName or persistence.cache.persistentVolumeClaim.storageClassName was set to "-". Setting "-" now renders storageClassName: "" to disable dynamic provisioning. Existing setups are unaffected: an empty value still omits the field (cluster default storage class) and an explicit class name is still rendered as-is.
Assets 4
Loading