Update instance guidance around larger instances and NUMA support

src/current/_includes/v25.2/prod-deployment/provision-cpu.md

@@ -1 +1 @@
-{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**32 vCPUs**{% endif %}
+{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**64 vCPUs**{% endif %}

src/current/_includes/v25.3/prod-deployment/provision-cpu.md

@@ -1 +1 @@
-{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**32 vCPUs**{% endif %}
+{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**64 vCPUs**{% endif %}

src/current/_includes/v25.4/prod-deployment/provision-cpu.md

@@ -1 +1 @@
-{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**32 vCPUs**{% endif %}
+{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**64 vCPUs**{% endif %}

src/current/_includes/v26.1/prod-deployment/provision-cpu.md

@@ -1 +1 @@
-{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**32 vCPUs**{% endif %}
+{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**64 vCPUs**{% endif %}

src/current/_includes/v26.2/prod-deployment/provision-cpu.md

@@ -1 +1 @@
-{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**32 vCPUs**{% endif %}
+{% if include.threshold == "absolute_minimum" %}**4 vCPUs**{% elsif include.threshold == "minimum" %}**8 vCPUs**{% elsif include.threshold == "maximum" %}**64 vCPUs**{% endif %}

src/current/v25.2/recommended-production-settings.md

@@ -70,6 +70,8 @@ In general, distribute your total vCPUs into the **largest possible nodes and sm
 
 - Cockroach Labs does not extensively test clusters with more than {% include {{ page.version.version }}/prod-deployment/provision-cpu.md threshold='maximum' %} per node. This is the recommended _maximum_ threshold.
 
+- CockroachDB should only run on single-NUMA instances. Running a CockroachDB process across NUMA nodes is not recommended and can heavily impact performance.
+
 ### Basic hardware recommendations
 
 After you [size your cluster](#sizing), you can determine the amount of RAM, storage capacity, and disk I/O from the number of vCPUs.

src/current/v25.3/recommended-production-settings.md

@@ -64,6 +64,8 @@ In general, distribute your total vCPUs into the **largest possible nodes and sm
 
 - Cockroach Labs does not extensively test clusters with more than {% include {{ page.version.version }}/prod-deployment/provision-cpu.md threshold='maximum' %} per node. This is the recommended _maximum_ threshold.
 
+- CockroachDB should only run on single-NUMA instances. Running a CockroachDB process across NUMA nodes is not recommended and can heavily impact performance.
+
 ### Basic hardware recommendations
 
 After you [size your cluster](#sizing), you can determine the amount of RAM, storage capacity, and disk I/O from the number of vCPUs.

src/current/v25.4/recommended-production-settings.md

@@ -64,6 +64,8 @@ In general, distribute your total vCPUs into the **largest possible nodes and sm
 
 - Cockroach Labs does not extensively test clusters with more than {% include {{ page.version.version }}/prod-deployment/provision-cpu.md threshold='maximum' %} per node. This is the recommended _maximum_ threshold.
 
+- CockroachDB should only run on single-NUMA instances. Running a CockroachDB process across NUMA nodes is not recommended and can heavily impact performance.
+
 ### Basic hardware recommendations
 
 After you [size your cluster](#sizing), you can determine the amount of RAM, storage capacity, and disk I/O from the number of vCPUs.

src/current/v26.1/recommended-production-settings.md

@@ -64,6 +64,8 @@ In general, distribute your total vCPUs into the **largest possible nodes and sm
 
 - Cockroach Labs does not extensively test clusters with more than {% include {{ page.version.version }}/prod-deployment/provision-cpu.md threshold='maximum' %} per node. This is the recommended _maximum_ threshold.
 
+- CockroachDB should only run on single-NUMA instances. Running a CockroachDB process across NUMA nodes is not recommended and can heavily impact performance.
+
 ### Basic hardware recommendations
 
 After you [size your cluster](#sizing), you can determine the amount of RAM, storage capacity, and disk I/O from the number of vCPUs.

src/current/v26.2/recommended-production-settings.md

@@ -64,6 +64,8 @@ In general, distribute your total vCPUs into the **largest possible nodes and sm
 
 - Cockroach Labs does not extensively test clusters with more than {% include {{ page.version.version }}/prod-deployment/provision-cpu.md threshold='maximum' %} per node. This is the recommended _maximum_ threshold.
 
+- CockroachDB should only run on single-NUMA instances. Running a CockroachDB process across NUMA nodes is not recommended and can heavily impact performance.
+
 ### Basic hardware recommendations
 
 After you [size your cluster](#sizing), you can determine the amount of RAM, storage capacity, and disk I/O from the number of vCPUs.