CRDB Operator docs improvements #22037
Conversation
✅ Deploy Preview for cockroachdb-api-docs canceled.
|
✅ Deploy Preview for cockroachdb-interactivetutorials-docs canceled.
|
| }, | ||
| { | ||
| "title": "Deploy in Kubernetes with CockroachDB Operator", | ||
| "title": "Deploy in Kubernetes with CockroachDB Operator (Recommended)", |
There was a problem hiding this comment.
Switching the order in the nav and adjusting the titles helps set the expectation that the CRDB operator is the "default" kubernetes deployment option. And for the older options, even while we can't explicitly label them as "legacy" or "deprecated" yet, we can implicitly label them as not preferred by calling them "other"
| The {{ site.data.products.cockroachdb-operator }} is a fully-featured Kubernetes operator that is designed for ease of deployment and scaling of multi-region clusters. To learn more, read the [{{ site.data.products.cockroachdb-operator }} documentation]({% link {{ site.versions["stable"] }}/cockroachdb-operator-overview.md %}). | ||
|
|
||
| New deployments of CockroachDB on Kubernetes are recommended to use the {{ site.data.products.cockroachdb-operator }}. To migrate an existing deployment to use the {{ site.data.products.cockroachdb-operator }}, read the [Helm]({% link {{ site.versions["stable"] }}/migrate-cockroachdb-kubernetes-helm.md %}) and [{{ site.data.products.public-operator }}]({% link {{ site.versions["stable"] }}/migrate-cockroachdb-kubernetes-operator.md %}) migration guides. | ||
| {{ site.data.alerts.callout_danger }} |
There was a problem hiding this comment.
This change turns the green "tip" note into a red "warning" note which catches the eye. Typically we reserve this for situations where guidance must be followed or a breakage will occur, but I think it's acceptable here considering that the more accurate "tip" callout has been insufficient.
✅ Netlify Preview
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
|
||
| The following deployment requirements and best practices apply to multi-region deployments with the {{ site.data.products.cockroachdb-operator }}: | ||
|
|
||
| - Plan to deploy one operator node per region. Each operator handles CockroachDB node management in its own region with no cross-region coordination, so you must factor this into your upgrade and maintenance strategy to ensure availability. |
There was a problem hiding this comment.
We could have operator deployment instead of node.
Each operator handles CockroachDB node management in its own region with no cross-region coordination
Not sure if this is the right wording because the operator in a region tries to connect to other region CockroachDB nodes and also builds join flag for a region.
There was a problem hiding this comment.
We could have operator deployment instead of node.
+1
because the operator in a region tries to connect to other region CockroachDB nodes
The operator does not quite do that, right? It's the cockroachdb process that connects to its instances in other regions.
There was a problem hiding this comment.
Sounds like it's just a matter of removing with no cross-region coordination because the main reader takeaway is just that should be multiple CRDB operator deployments and they need to plan upgrades accordingly.
pritesh-lahoti
left a comment
pritesh-lahoti
left a comment
There was a problem hiding this comment.
LGTM. Pre-approving
|
|
||
| The following deployment requirements and best practices apply to multi-region deployments with the {{ site.data.products.cockroachdb-operator }}: | ||
|
|
||
| - Plan to deploy one operator node per region. Each operator handles CockroachDB node management in its own region with no cross-region coordination, so you must factor this into your upgrade and maintenance strategy to ensure availability. |
There was a problem hiding this comment.
We could have operator deployment instead of node.
+1
because the operator in a region tries to connect to other region CockroachDB nodes
The operator does not quite do that, right? It's the cockroachdb process that connects to its instances in other regions.
3 participants
(Before) Left hand navigation:
(Before) Boilerplate note on old Kubernetes pages:
(After) Left hand navigation:
(After) Boilerplate note on old Kubernetes pages:
TODO: Backport to other supported versions once content is approved