Skip to content

Commit 10511b8

Browse files
committed
Add release notes for #127
1 parent e0fff39 commit 10511b8

1 file changed

Lines changed: 42 additions & 0 deletions

File tree

Lines changed: 42 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,42 @@
1+
# Release Notes for Issue #127: Expose tolerations and nodeSelector in chart values
2+
3+
## New Feature
4+
5+
### What Changed
6+
`tolerations` and `nodeSelector` are now declared in
7+
`charts/restate-operator-helm/values.yaml`. The `tolerations` template was
8+
already wired in `templates/deployment.yaml`; this just makes it discoverable
9+
via the values file. A matching `nodeSelector` template block was added.
10+
11+
### Why This Matters
12+
On clusters with node taints — per-workload taint pools, dedicated controller
13+
node groups, Karpenter-style provisioners — the operator pod previously had no
14+
way to schedule through Helm values alone. Users had to fork the chart or layer
15+
a kustomize+helm post-render patch. Declaring these standard pod scheduling
16+
fields in `values.yaml` brings the chart in line with the typical operator
17+
chart convention (cert-manager, external-secrets, AWS Load Balancer Controller,
18+
etc.).
19+
20+
### Impact on Users
21+
- Existing deployments: No impact, both fields are optional
22+
- New deployments: Can pin the operator pod to specific nodes or tolerate
23+
taints through Helm values, instead of forking the chart or post-rendering
24+
with kustomize
25+
26+
### Migration Guidance
27+
No migration required. To use the new fields:
28+
29+
30+
```yaml
31+
tolerations:
32+
- key: "service"
33+
operator: "Exists"
34+
effect: "NoSchedule"
35+
36+
nodeSelector:
37+
workload: controllers
38+
```
39+
40+
41+
### Related Issues
42+
- Issue #127: Expose tolerations / nodeSelector in restate-operator-helm chart

0 commit comments

Comments
 (0)