Cherry pick #3429 and #3431; longevity results and GKE annotation fix (#3437)

* Add 2.0 longevity results (#3429)

Adding longevity results for 2.0. These results are not complete and may be inaccurate. The tests were stopped early and the teardown scripts/functions did not collect everything properly.

* Use proper internal GKE service annotation (#3431)

Problem: The service annotation for the GKE services was improperly formatted.

Solution: Fix the annotation so it provisions the proper service.

Author: sjberman

charts/nginx-gateway-fabric/values.yaml

@@ -493,7 +493,7 @@ gateways: []
 #   gatewayClassName: nginx
 #   infrastructure:
 #     annotations:
-#       service.annotations.networking.gke.io/load-balancer-type: Internal
+#       networking.gke.io/load-balancer-type: Internal
 #   listeners:
 #     - name: https
 #       port: 80

tests/results/longevity/2.0.0/oss-cpu.png

@@ -0,0 +1,43 @@
+# Results
+
+## Summary
+
+These results are incomplete and may be inaccurate. The test was stopped manually about 1 day early, and results such as logs and traffic summary were not collected properly by the teardown scripts/functions.
+
+There are fewer dashboards to collect now after the change in architecture, as we don't have the same metrics as we did before, mainly relating to reloads.
+
+One thing of note is the significant increase in memory usage for the NGF control plane container. For our NGINX Plus run, the opposite occurred, where the NGINX container had high memory.
+
+## Test environment
+
+NGINX Plus: false
+
+NGINX Gateway Fabric:
+
+- Commit: cc3c907ff668d886cac719df2d77b685370ad5f8
+- Date: 2025-05-30T18:25:58Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 3
+- k8s version: v1.32.4-gke.1106006
+- vCPUs per node: 2
+- RAM per node: 4015484Ki
+- Max pods per node: 110
+- Zone: us-west2-a
+- Instance Type: e2-medium
+
+## Key Metrics
+
+### Containers memory
+
+![oss-memory.png](oss-memory.png)
+
+### NGF Container Memory
+
+![oss-ngf-memory.png](oss-ngf-memory.png)
+
+### Containers CPU
+
+![oss-cpu.png](oss-cpu.png)

tests/results/longevity/2.0.0/oss-memory.png

@@ -0,0 +1,65 @@
+# Results
+
+## Summary
+
+These results are incomplete and may be inaccurate. The test was stopped manually about 1 day early, and results such as logs and traffic summary were not collected properly by the teardown scripts/functions.
+
+There are fewer dashboards to collect now after the change in architecture, as we don't have the same metrics as we did before, mainly relating to reloads.
+
+One thing of note is the significant increase in memory usage for the NGINX container.
+
+## Test environment
+
+NGINX Plus: true
+
+NGINX Gateway Fabric:
+
+- Commit: cc3c907ff668d886cac719df2d77b685370ad5f8
+- Date: 2025-05-30T18:25:58Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 3
+- k8s version: v1.32.4-gke.1106006
+- vCPUs per node: 2
+- RAM per node: 4015484Ki
+- Max pods per node: 110
+- Zone: us-west2-a
+- Instance Type: e2-medium
+
+## Traffic
+
+HTTP:
+
+```text
+```
+
+HTTPS:
+
+```text
+```
+
+
+## Error Logs
+
+### nginx-gateway
+
+
+
+### nginx
+2025/06/01 15:34:12 [error] 78#78: *157671523 no live upstreams while connecting to upstream, client: 35.236.69.111, server: cafe.example.com, request: "GET /tea HTTP/1.1", upstream: "http://longevity_tea_80/tea", host: "cafe.example.com"
+
+## Key Metrics
+
+### Containers memory
+
+![plus-memory.png](plus-memory.png)
+
+### NGF Container Memory
+
+![plus-ngf-memory.png](plus-ngf-memory.png)
+
+### Containers CPU
+
+![plus-cpu.png](plus-cpu.png)

tests/results/longevity/2.0.0/oss-ngf-memory.png

@@ -6,7 +6,7 @@ spec:
   gatewayClassName: nginx
   infrastructure:
     annotations:
-      service.annotations.networking.gke.io/load-balancer-type: Internal
+      networking.gke.io/load-balancer-type: Internal
   listeners:
   - name: http
     port: 80