nginx · sjberman · Dec 16, 2025 · Dec 12, 2025 · Dec 15, 2025 · Dec 16, 2025
@@ -0,0 +1,90 @@
+# Results
+
+## Test environment
+
+NGINX Plus: false
+
+NGINX Gateway Fabric:
+
+- Commit: 89aee48bf6e660a828ffd32ca35fc7f52e358e00
+- Date: 2025-12-12T20:04:38Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 12
+- k8s version: v1.33.5-gke.1308000
+- vCPUs per node: 16
+- RAM per node: 65851520Ki
+- Max pods per node: 110
+- Zone: us-west1-b
+- Instance Type: n2d-standard-16
+
+## Summary: 
+
+- Latency continues to grow slightly, per the trend of past releases.
+
+## Test1: Running latte path based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.03, 999.99
+Duration      [total, attack, wait]             30s, 29.999s, 991.978µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  816.445µs, 1.069ms, 1.045ms, 1.166ms, 1.217ms, 1.385ms, 23.061ms
+Bytes In      [total, mean]                     4740000, 158.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test2: Running coffee header based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.04, 1000.00
+Duration      [total, attack, wait]             30s, 29.999s, 1.132ms
+Latencies     [min, mean, 50, 90, 95, 99, max]  840.624µs, 1.096ms, 1.073ms, 1.204ms, 1.26ms, 1.44ms, 16.79ms
+Bytes In      [total, mean]                     4770000, 159.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test3: Running coffee query based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.04, 1000.00
+Duration      [total, attack, wait]             30s, 29.999s, 1.067ms
+Latencies     [min, mean, 50, 90, 95, 99, max]  825.3µs, 1.095ms, 1.071ms, 1.201ms, 1.256ms, 1.444ms, 16.845ms
+Bytes In      [total, mean]                     5010000, 167.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test4: Running tea GET method based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.02, 999.99
+Duration      [total, attack, wait]             30s, 29.999s, 954.141µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  818.006µs, 1.079ms, 1.059ms, 1.187ms, 1.241ms, 1.411ms, 14.873ms
+Bytes In      [total, mean]                     4680000, 156.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test5: Running tea POST method based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.03, 1000.00
+Duration      [total, attack, wait]             30s, 29.999s, 992.607µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  808.16µs, 1.086ms, 1.064ms, 1.196ms, 1.248ms, 1.42ms, 17.019ms
+Bytes In      [total, mean]                     4680000, 156.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
@@ -0,0 +1,90 @@
+# Results
+
+## Test environment
+
+NGINX Plus: true
+
+NGINX Gateway Fabric:
+
+- Commit: 89aee48bf6e660a828ffd32ca35fc7f52e358e00
+- Date: 2025-12-12T20:04:38Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 12
+- k8s version: v1.33.5-gke.1308000
+- vCPUs per node: 16
+- RAM per node: 65851520Ki
+- Max pods per node: 110
+- Zone: us-west1-b
+- Instance Type: n2d-standard-16
+
+## Summary: 
+
+- Latency looks to have improved slightly.
+
+## Test1: Running latte path based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.04, 1000.01
+Duration      [total, attack, wait]             30s, 29.999s, 880.439µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  691.14µs, 886.932µs, 867.964µs, 976.348µs, 1.018ms, 1.153ms, 10.358ms
+Bytes In      [total, mean]                     4830000, 161.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test2: Running coffee header based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.04, 1000.01
+Duration      [total, attack, wait]             30s, 29.999s, 923.361µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  726.599µs, 948.386µs, 919.848µs, 1.025ms, 1.07ms, 1.262ms, 22.38ms
+Bytes In      [total, mean]                     4860000, 162.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test3: Running coffee query based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.04, 1000.01
+Duration      [total, attack, wait]             30s, 29.999s, 980.118µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  741.198µs, 949.099µs, 920.511µs, 1.025ms, 1.067ms, 1.241ms, 19.154ms
+Bytes In      [total, mean]                     5100000, 170.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test4: Running tea GET method based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.01, 999.98
+Duration      [total, attack, wait]             30.001s, 30s, 997.667µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  716.164µs, 903.954µs, 881.394µs, 978.714µs, 1.019ms, 1.192ms, 21.825ms
+Bytes In      [total, mean]                     4770000, 159.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
+
+## Test5: Running tea POST method based routing
+
+```text
+Requests      [total, rate, throughput]         30000, 1000.01, 999.97
+Duration      [total, attack, wait]             30.001s, 30s, 919.688µs
+Latencies     [min, mean, 50, 90, 95, 99, max]  708.879µs, 925.517µs, 903.767µs, 1.012ms, 1.054ms, 1.21ms, 22.009ms
+Bytes In      [total, mean]                     4770000, 159.00
+Bytes Out     [total, mean]                     0, 0.00
+Success       [ratio]                           100.00%
+Status Codes  [code:count]                      200:30000  
+Error Set:
+```
@@ -0,0 +1,83 @@
+# Results
+
+## Test environment
+
+NGINX Plus: false
+
+NGINX Gateway Fabric:
+
+- Commit: 89aee48bf6e660a828ffd32ca35fc7f52e358e00
+- Date: 2025-12-12T20:04:38Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 3
+- k8s version: v1.33.5-gke.1308000
+- vCPUs per node: 2
+- RAM per node: 4015672Ki
+- Max pods per node: 110
+- Zone: us-west2-a
+- Instance Type: e2-medium
+
+## Summary:
+
+- Still a lot of non-2xx or 3xx responses, many more than last time. Socket errors are all mostly read errors, with no write errors and fewer timeout errors.
+- We observe a continual increase in NGINX memory usage over time which could indicate a memory leak. Will bring this up with the Agent team.
+- CPU usage remained consistent with past results.
+- Error contacting TokenReview API, but may be a one-off.
+
+## Traffic
+
+HTTP:
+
+```text
+Running 5760m test @ http://cafe.example.com/coffee
+  2 threads and 100 connections
+  Thread Stats   Avg      Stdev     Max   +/- Stdev
+    Latency   190.35ms  141.74ms   2.00s    83.52%
+    Req/Sec   289.84    187.59     3.52k    63.68%
+  195509968 requests in 5760.00m, 66.75GB read
+  Socket errors: connect 0, read 315485, write 0, timeout 6584
+  Non-2xx or 3xx responses: 1763516
+Requests/sec:    565.71
+Transfer/sec:    202.53KB
+```
+
+HTTPS:
+
+```text
+Running 5760m test @ https://cafe.example.com/tea
+  2 threads and 100 connections
+  Thread Stats   Avg      Stdev     Max   +/- Stdev
+    Latency   180.03ms  106.92ms   1.94s    67.25%
+    Req/Sec   287.34    184.95     1.73k    63.36%
+  193842103 requests in 5760.00m, 65.22GB read
+  Socket errors: connect 0, read 309621, write 0, timeout 1
+Requests/sec:    560.89
+Transfer/sec:    197.88KB
+```
+## Key Metrics
+
+### Containers memory
+
+![oss-memory.png](oss-memory.png)
+
+### Containers CPU
+
+![oss-cpu.png](oss-cpu.png)
+
+## Error Logs
+
+### nginx-gateway
+
+error=rpc error: code = Internal desc = error creating TokenReview: context canceled;level=error;logger=agentGRPCServer;msg=error validating connection;stacktrace=github.com/nginx/nginx-gateway-fabric/v2/internal/controller/nginx/agent/grpc/interceptor.(*ContextSetter).Stream.ContextSetter.Stream.func1
+	/opt/actions-runner/_work/nginx-gateway-fabric/nginx-gateway-fabric/internal/controller/nginx/agent/grpc/interceptor/interceptor.go:62
+google.golang.org/grpc.(*Server).processStreamingRPC
+	/opt/actions-runner/_work/nginx-gateway-fabric/nginx-gateway-fabric/.gocache/google.golang.org/[email protected]/server.go:1721
+google.golang.org/grpc.(*Server).handleStream
+	/opt/actions-runner/_work/nginx-gateway-fabric/nginx-gateway-fabric/.gocache/google.golang.org/[email protected]/server.go:1836
+google.golang.org/grpc.(*Server).serveStreams.func2.1
+	/opt/actions-runner/_work/nginx-gateway-fabric/nginx-gateway-fabric/.gocache/google.golang.org/[email protected]/server.go:1063;ts=2025-12-16T17:35:17Z
+
+### nginx
@@ -0,0 +1,83 @@
+# Results
+
+## Test environment
+
+NGINX Plus: false
+
+NGINX Gateway Fabric:
+
+- Commit: 89aee48bf6e660a828ffd32ca35fc7f52e358e00
+- Date: 2025-12-12T20:04:38Z
+- Dirty: false
+
+GKE Cluster:
+
+- Node count: 3
+- k8s version: v1.33.5-gke.1308000
+- vCPUs per node: 2
+- RAM per node: 4015672Ki
+- Max pods per node: 110
+- Zone: us-west2-a
+- Instance Type: e2-medium
+
+## Summary:
+
+- Consistent traffic results from 2.2.
+- We observe a continual increase in NGINX memory usage over time which could indicate a memory leak. Will bring this up with the Agent team.
+- CPU usage remained consistent with past results.
+- Still get some "no live upstreams" errors.
+
+## Traffic
+
+HTTP:
+
+```text
+Running 5760m test @ http://cafe.example.com/coffee
+  2 threads and 100 connections
+  Thread Stats   Avg      Stdev     Max   +/- Stdev
+    Latency   184.82ms  102.91ms   1.45s    65.52%
+    Req/Sec   284.19    179.74     1.52k    63.62%
+  192198367 requests in 5760.00m, 65.91GB read
+  Socket errors: connect 0, read 0, write 0, timeout 108
+  Non-2xx or 3xx responses: 5
+Requests/sec:    556.13
+Transfer/sec:    199.96KB
+```
+
+HTTPS:
+
+```text
+Running 5760m test @ https://cafe.example.com/tea
+  2 threads and 100 connections
+  Thread Stats   Avg      Stdev     Max   +/- Stdev
+    Latency   185.02ms  102.92ms   1.50s    65.52%
+    Req/Sec   283.70    179.19     1.43k    63.75%
+  191866398 requests in 5760.00m, 64.73GB read
+  Socket errors: connect 0, read 0, write 0, timeout 114
+  Non-2xx or 3xx responses: 6
+Requests/sec:    555.17
+Transfer/sec:    196.40KB
+```
+## Key Metrics
+
+### Containers memory
+
+![oss-memory.png](oss-memory.png)
+
+### Containers CPU
+
+![oss-cpu.png](oss-cpu.png)
+
+## Error Logs
+
+### nginx-gateway
+
+### nginx
+
+
+
+
+10.168.0.90 - - [16/Dec/2025:15:47:08 +0000] "GET /tea HTTP/1.1" 502 150 "-" "-"
+2025/12/16 15:47:08 [error] 26#26: *361983622 no live upstreams while connecting to upstream, client: 10.168.0.90, server: cafe.example.com, request: "GET /tea HTTP/1.1", upstream: "http://longevity_tea_80/tea", host: "cafe.example.com"
+10.168.0.90 - - [16/Dec/2025:12:49:07 +0000] "GET /coffee HTTP/1.1" 502 150 "-" "-"
+2025/12/16 12:49:07 [error] 25#25: *350621339 no live upstreams while connecting to upstream, client: 10.168.0.90, server: cafe.example.com, request: "GET /coffee HTTP/1.1", upstream: "http://longevity_coffee_80/coffee", host: "cafe.example.com"