Add support for tolerations

[amdove]

The helm chart does not allow for adding tolerations to the deployment objects. When run on default node groups, we're seeing high resource usage, impacting other deployments. To remediate this, we'd like to use Karpenter to assign Tonic Textual pods to a separate node pool to isolate any impact from other running workloads. Without the ability to assign tolerations to the pods deployed via the helm chart, we're unable to use this strategy.