Changes and fixes before the release (docs/small tweaks) (#750)

- update README.md with new docker instructions
- update docs/configuration.md
- update .github/workflows to have pinned action versions
- gofmt events package
- fix small bugs in CI scripts
- reduce config options for internal/perf/monitor and config. A ring buffer is used to keep 1hr of entries at max 5s granularity. For long term stats use prometheus monitoring on /metrics

Fixes #744

.github/workflows/closeinactive.yml

@@ -11,7 +11,7 @@ jobs:
       issues: write
       pull-requests: write
     steps:
-      - uses: actions/stale@v9
+      - uses: actions/stale@b5d41d4e1d5dceea10e7104786b73624c18a190f #v10.2.0
         with:
           days-before-issue-stale: 14
           days-before-issue-close: 14

.github/workflows/config-schema.yml

@@ -21,7 +21,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
       - name: Validate JSON Schema
         run: |
@@ -45,7 +45,7 @@ jobs:
           echo "✓ config-schema.json is valid"
 
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 #v6.2.0
         with:
           python-version: "3.x"
 

.github/workflows/containers.yml

@@ -38,7 +38,7 @@ jobs:
       fail-fast: false
     steps:
       - name: Checkout code
-        uses: actions/checkout@v6
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
       - name: Free up disk space
         if: matrix.platform == 'rocm'
@@ -58,13 +58,13 @@ jobs:
       # no-op for amd64-only builds, so leaving it on for every matrix
       # entry keeps the workflow simple.
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v4
+        uses: docker/setup-qemu-action@ce360397dd3f832beb865e1373c09c0e9f86d70a #v4.0.0
 
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v4
+        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd #v4.0.0
 
       - name: Log in to GitHub Container Registry
-        uses: docker/login-action@v4
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121 #v4.1.0
         with:
           registry: ghcr.io
           username: ${{ github.actor }}

.github/workflows/go-ci-windows.yml

@@ -31,17 +31,17 @@ jobs:
   run-tests:
     runs-on: windows-latest
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
     - name: Set up Go
-      uses: actions/setup-go@v4
+      uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c #6.4.0
       with:
-        go-version: '1.23'
+        go-version-file: go.mod
 
     # cache simple-responder to save the build time
     - name: Restore Simple Responder
       id: restore-simple-responder
-      uses: actions/cache/restore@v4
+      uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae #v5.0.5
       with:
         path: ./build
         key: ${{ runner.os }}-simple-responder-${{ hashFiles('cmd/simple-responder/simple-responder.go') }}
@@ -56,7 +56,7 @@ jobs:
       # nothing new to save ... skip this step
       if: steps.restore-simple-responder.outputs.cache-hit != 'true'
       id: save-simple-responder
-      uses: actions/cache/save@v4
+      uses: actions/cache/save@27d5ce7f107fe9357f9df03efb73ab90386fccae #v5.0.5
       with:
         path: ./build
         key: ${{ runner.os }}-simple-responder-${{ hashFiles('cmd/simple-responder/simple-responder.go') }}

.github/workflows/go-ci.yml

@@ -30,24 +30,24 @@ jobs:
   run-tests:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
       - name: Set up Go
-        uses: actions/setup-go@v4
+        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c #6.4.0
         with:
           go-version-file: go.mod
 
       # Only run in this linux based runner
       - name: Check Formatting
         run: |
-          if [ "$(gofmt -l . | grep -v 'event/.*_test.go' | wc -l)" -gt 0 ]; then
-            gofmt -l . | grep -v 'event/.*_test.go'
+          if [ "$(gofmt -l . | wc -l)" -gt 0 ]; then
+            gofmt -l .
             exit 1
           fi
       # cache simple-responder to save the build time
       - name: Restore Simple Responder
         id: restore-simple-responder
-        uses: actions/cache/restore@v4
+        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae #v5.0.5
         with:
           path: ./build
           key: ${{ runner.os }}-simple-responder-${{ hashFiles('cmd/simple-responder/simple-responder.go') }}
@@ -60,7 +60,7 @@ jobs:
         # nothing new to save ... skip this step
         if: steps.restore-simple-responder.outputs.cache-hit != 'true'
         id: save-simple-responder
-        uses: actions/cache/save@v4
+        uses: actions/cache/save@27d5ce7f107fe9357f9df03efb73ab90386fccae #v5.0.5
         with:
           path: ./build
           key: ${{ runner.os }}-simple-responder-${{ hashFiles('cmd/simple-responder/simple-responder.go') }}

.github/workflows/release.yml

@@ -20,14 +20,14 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@v4
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
         with:
           fetch-depth: 0
           ref: ${{ github.event.inputs.tag || github.ref }}
       - name: Set up Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c #6.4.0
       - name: Set up Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@48b55a011bda9f5d6aeb4c2d9c7362e8dae4041e # 6.4.0
         with:
           node-version: "24"
       - name: Install dependencies and build UI
@@ -37,7 +37,7 @@ jobs:
           npm run build
 
       - name: Run GoReleaser
-        uses: goreleaser/goreleaser-action@v6
+        uses: goreleaser/goreleaser-action@1a80836c5c9d9e5755a25cb59ec6f45a3b5f41a8 #7.2.1
         with:
           # either 'goreleaser' (default) or 'goreleaser-pro'
           distribution: goreleaser
@@ -61,7 +61,7 @@ jobs:
           fi
 
       - name: "Trigger tap repository update"
-        uses: peter-evans/repository-dispatch@v2
+        uses: peter-evans/repository-dispatch@28959ce8df70de7be546dd1250a005dd32156697 #4.0.1
         with:
           token: ${{ secrets.TAP_REPO_PAT }}
           repository: mostlygeek/homebrew-llama-swap

.github/workflows/ui-tests.yml

@@ -20,10 +20,10 @@ jobs:
   run-tests:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@v4
+    - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
     - name: Set up Node.js
-      uses: actions/setup-node@v4
+      uses: actions/setup-node@48b55a011bda9f5d6aeb4c2d9c7362e8dae4041e # 6.4.0
       with:
         node-version: '24'
         cache: 'npm'

.github/workflows/unified-docker.yml

@@ -75,7 +75,7 @@ jobs:
         backend: ${{ fromJSON(needs.setup.outputs.matrix) }}
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # 6.0.2
 
       - name: Free up disk space
         run: |
@@ -94,11 +94,11 @@ jobs:
       # llama-swap-builder (which has ccache warm) to avoid exhausting disk.
       - name: Set up Docker Buildx
         if: ${{ !env.ACT }}
-        uses: docker/setup-buildx-action@v3
+        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd #v4.0.0
 
       - name: Log in to GitHub Container Registry
         if: ${{ !env.ACT }}
-        uses: docker/login-action@v3
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121 #v4.1.0
         with:
           registry: ghcr.io
           username: ${{ github.actor }}

README.md

@@ -57,8 +57,9 @@ Built in Go for performance and simplicity, llama-swap has zero dependencies and
 - ✅ Customizable
   - Run concurrent models with a custom DSL swap matrix ([#643](https://github.com/mostlygeek/llama-swap/issues/643))
   - Automatic unloading of models after timeout by setting a `ttl`
-  - Reliable Docker and Podman support using `cmd` and `cmdStop` together
+  - Docker and Podman support using `cmd` and `cmdStop` together
   - Preload models on startup with `hooks` ([#235](https://github.com/mostlygeek/llama-swap/pull/235))
+  - Apply filters to requests to control inference with `stripParams`, `setParams` and `setParamsByID`
 
 ### Web UI
 
@@ -94,8 +95,24 @@ llama-swap can be installed in multiple ways
 
 ### Docker Install ([download images](https://github.com/mostlygeek/llama-swap/pkgs/container/llama-swap))
 
-Nightly container images with llama-swap and llama-server are built for multiple platforms (cuda, vulkan, intel, etc.) including [non-root variants with improved security](docs/container-security.md).
-The stable-diffusion.cpp server is also included for the musa and vulkan platforms.
+Two types of container images are built nightly for llama-swap:
+
+1. A unified container with llama-server, ik-llama-server, stable-diffusion.cpp, whisper.cpp and llama-swap built from source. This is only available for cuda and vulkan but has more capabilities. This one is recommended for use.
+2. A legacy image that is based on llama.cpp's images and llama-swap copied into the container. Use this one if you prefer to stay close to llama.cpp's container images.
+
+#### Unified container (Recommended)
+
+```shell
+$ docker pull ghcr.io/mostlygeek/llama-swap:unified-cuda
+
+# run with a custom configuration and models directory
+$ docker run -it --rm --runtime nvidia -p 9292:8080 \
+ -v /path/to/models:/models \
+ -v /path/to/custom/config.yaml:/etc/llama-swap/config/config.yaml \
+ ghcr.io/mostlygeek/llama-swap:unified-cuda
+```
+
+#### Legacy container
 
 ```shell
 $ docker pull ghcr.io/mostlygeek/llama-swap:cuda
@@ -105,14 +122,6 @@ $ docker run -it --rm --runtime nvidia -p 9292:8080 \
  -v /path/to/models:/models \
  -v /path/to/custom/config.yaml:/app/config.yaml \
  ghcr.io/mostlygeek/llama-swap:cuda
-
-# configuration hot reload supported with a
-# directory volume mount
-$ docker run -it --rm --runtime nvidia -p 9292:8080 \
- -v /path/to/models:/models \
- -v /path/to/custom/config.yaml:/app/config.yaml \
- -v /path/to/config:/config \
- ghcr.io/mostlygeek/llama-swap:cuda -config /config/config.yaml -watch-config
 ```
 
 <details>
@@ -268,6 +277,6 @@ For Python based inference servers like vllm or tabbyAPI it is recommended to ru
 ## Star History
 
 > [!NOTE]
-> ⭐️ Star this project to help others discover it!
+> Thank you to everyone who has given this project a ⭐️!
 
 [![Star History Chart](https://api.star-history.com/svg?repos=mostlygeek/llama-swap&type=Date)](https://www.star-history.com/#mostlygeek/llama-swap&Date)

cmd/monitor-test/main.go

@@ -75,7 +75,7 @@ func main() {
 	}
 
 	if *stream {
-		m, _ := perf.New(config.PerformanceConfig{Enable: true, Every: every}, l)
+		m, _ := perf.New(config.PerformanceConfig{Every: every}, l)
 		m.Start()
 		defer m.Stop()
 		sysCh, gpuCh, unsub := m.Subscribe()

config-schema.json

@@ -145,33 +145,21 @@
         "performance": {
             "type": "object",
             "properties": {
-                "enable": {
+                "disabled": {
                     "type": "boolean",
-                    "default": true,
-                    "description": "Enable or disable system performance monitoring."
+                    "default": false,
+                    "description": "Disable system performance monitoring."
                 },
                 "every": {
                     "type": "string",
                     "pattern": "^[-+]?(\\d+(\\.\\d+)?(ns|us|ms|s|m|h))+$",
                     "default": "15s",
                     "description": "Delay between polling for new performance statistics. Minimum duration is 1s. Lower values use more RAM as stats are kept in memory."
-                },
-                "maxAge": {
-                    "type": "string",
-                    "pattern": "^[-+]?(\\d+(\\.\\d+)?(ns|us|ms|s|m|h))+$",
-                    "default": "1h",
-                    "description": "Maximum age of performance statistics before they are eligible for garbage collection."
-                },
-                "gc": {
-                    "type": "string",
-                    "pattern": "^[-+]?(\\d+(\\.\\d+)?(ns|us|ms|s|m|h))+$",
-                    "default": "5m",
-                    "description": "Garbage collection frequency for clearing old performance statistics."
                 }
             },
             "additionalProperties": false,
             "default": {},
-            "description": "Configuration for system monitoring statistics. Timing values are duration strings like 1s, 1h30m, 90m, 2h10s."
+            "description": "Configuration for CPU, RAM and GPU monitoring statistics."
         },
         "startPort": {
             "type": "integer",

config.example.yaml

@@ -58,25 +58,15 @@ captureBuffer: 15
 # performance: configuration for system monitoring statistics
 # - timing values are duration strings like 1s, 1h30m, 90m, 2h10s, etc.
 performance:
-  # enabled: boolean
-  # - default: true
-  enable: true
+  # disabled: boolean
+  # - default: false
+  disabled: false
 
   # every: delay between polling for new performance statistics
-  # - default: 15s
-  # - minimum duration 1s
-  # - note: setting this very low will use up more RAM as stats are kept in memory.
+  # - default: 5s
+  # - minimum duration 5s
   every: 15s
 
-  # maxAge: maximum age of a performance statistics before it is eligible for garbage collection
-  # - default: 1h
-  maxAge: 12h
-
-  # gc: garbage collection frequency in seconds
-  # - how many seconds the garbage collector runs to clear old stats
-  # - default 5m
-  gc: 5m
-
 # startPort: sets the starting port number for the automatic ${PORT} macro.
 # - optional, default: 5800
 # - the ${PORT} macro can be used in model.cmd and model.proxy settings
@@ -118,8 +108,7 @@ globalTTL: 0
 macros:
   # Example of a multi-line macro
   "latest-llama": >
-    /path/to/llama-server/llama-server-ec9e0301
-    --port ${PORT}
+    /path/to/llama-server/llama-server-ec9e0301 --port ${PORT}
 
   "default_ctx": 4096
 
@@ -279,7 +268,8 @@ models:
 
       # the ${temp} macro will remain a float
       temperature: ${temp}
-      note: "The ${MODEL_ID} is running on port ${PORT} temp=${temp}, context=${default_ctx}"
+      note: "The ${MODEL_ID} is running on port ${PORT} temp=${temp},
+        context=${default_ctx}"
 
       a_list:
         - 1
@@ -291,7 +281,7 @@ models:
         b: 2
         # objects can contain complex types with macro substitution
         # becomes: c: [0.7, false, "model: llama"]
-        c: ["${temp}", false, "model: ${MODEL_ID}"]
+        c: [ "${temp}", false, "model: ${MODEL_ID}" ]
 
     # concurrencyLimit: overrides the allowed number of active parallel requests to a model
     # - optional, default: 0

docs/configuration.md

@@ -146,6 +146,18 @@ metricsMaxInMemory: 1000
 # - set to 0 to disable
 captureBuffer: 15
 
+# performance: configuration for system monitoring statistics
+# - timing values are duration strings like 1s, 1h30m, 90m, 2h10s, etc.
+performance:
+  # disabled: boolean
+  # - default: false
+  enable: true
+
+  # every: delay between polling for new performance statistics
+  # - default: 5s
+  # - minimum duration 5s
+  every: 5s
+
 # startPort: sets the starting port number for the automatic ${PORT} macro.
 # - optional, default: 5800
 # - the ${PORT} macro can be used in model.cmd and model.proxy settings
@@ -187,8 +199,7 @@ globalTTL: 0
 macros:
   # Example of a multi-line macro
   "latest-llama": >
-    /path/to/llama-server/llama-server-ec9e0301
-    --port ${PORT}
+    /path/to/llama-server/llama-server-ec9e0301 --port ${PORT}
 
   "default_ctx": 4096
 
@@ -348,7 +359,8 @@ models:
 
       # the ${temp} macro will remain a float
       temperature: ${temp}
-      note: "The ${MODEL_ID} is running on port ${PORT} temp=${temp}, context=${default_ctx}"
+      note: "The ${MODEL_ID} is running on port ${PORT} temp=${temp},
+        context=${default_ctx}"
 
       a_list:
         - 1

internal/perf/monitor.go

@@ -31,7 +31,7 @@ type Monitor struct {
 }
 
 func ringCapacity(c config.PerformanceConfig) int {
-	n := int(c.MaxAge / c.Every)
+	n := int(time.Hour / c.Every)
 	if n < 1 {
 		n = 1
 	}
@@ -43,12 +43,6 @@ func New(c config.PerformanceConfig, logger *logmon.Monitor) (*Monitor, error) {
 	if c.Every < 100*time.Millisecond {
 		c.Every = 100 * time.Millisecond
 	}
-	if c.GC < 1*time.Second {
-		c.GC = 1 * time.Second
-	}
-	if c.MaxAge < 1*time.Minute {
-		c.MaxAge = 1 * time.Minute
-	}
 
 	if logger == nil {
 		return nil, errors.New("logger is required")
@@ -92,7 +86,9 @@ func (m *Monitor) UpdateConfig(newConf config.PerformanceConfig) {
 	m.sysRing = ring.NewBuffer[SysStat](capacity)
 	m.gpuRing = ring.NewBuffer[[]GpuStat](capacity)
 	m.mutex.Unlock()
-	m.Start()
+	if !newConf.Disabled {
+		m.Start()
+	}
 }
 
 // Subscribe returns channels to listen to system and GPU stats.

internal/perf/monitor_test.go

@@ -24,26 +24,19 @@ func TestNew_DefaultConfig(t *testing.T) {
 	require.NotNil(t, m)
 
 	assert.Equal(t, 100*time.Millisecond, m.conf.Every)
-	assert.Equal(t, 1*time.Second, m.conf.GC)
-	assert.Equal(t, 1*time.Minute, m.conf.MaxAge)
 }
 
 func TestNew_CustomConfig(t *testing.T) {
 	logger := newTestLogger()
 
 	cfg := config.PerformanceConfig{
-		Enable: true,
-		Every:  500 * time.Millisecond,
-		GC:     5 * time.Second,
-		MaxAge: 10 * time.Minute,
+		Every: 500 * time.Millisecond,
 	}
 
 	m, err := New(cfg, logger)
 	require.NoError(t, err)
 
 	assert.Equal(t, 500*time.Millisecond, m.conf.Every)
-	assert.Equal(t, 5*time.Second, m.conf.GC)
-	assert.Equal(t, 10*time.Minute, m.conf.MaxAge)
 }
 
 func TestNew_NilLogger(t *testing.T) {
@@ -56,18 +49,13 @@ func TestNew_BelowMinimumConfig(t *testing.T) {
 	logger := newTestLogger()
 
 	cfg := config.PerformanceConfig{
-		Enable: true,
-		Every:  1 * time.Millisecond,
-		GC:     100 * time.Millisecond,
-		MaxAge: 1 * time.Second,
+		Every: 1 * time.Millisecond,
 	}
 
 	m, err := New(cfg, logger)
 	require.NoError(t, err)
 
 	assert.Equal(t, 100*time.Millisecond, m.conf.Every)
-	assert.Equal(t, 1*time.Second, m.conf.GC)
-	assert.Equal(t, 1*time.Minute, m.conf.MaxAge)
 }
 
 func TestSubscribe_ReturnsChannels(t *testing.T) {

llama-swap.go

@@ -93,12 +93,17 @@ func main() {
 		listenStr = &defaultPort
 	}
 
-	mon, err := perf.New(conf.Performance, mainLogger)
-	if err != nil {
-		mainLogger.Errorf("failed to create monitor: %s", err.Error())
-		os.Exit(1)
+	var mon *perf.Monitor
+	if !conf.Performance.Disabled {
+		mon, err = perf.New(conf.Performance, mainLogger)
+		if err != nil {
+			mainLogger.Errorf("failed to create monitor: %s", err.Error())
+			os.Exit(1)
+		}
+		mon.Start()
+	} else {
+		mainLogger.Info("performance monitoring is disabled")
 	}
-	mon.Start()
 
 	// Setup channels for server management
 	exitChan := make(chan struct{})
@@ -108,7 +113,7 @@ func main() {
 	// Context that bounds the lifetime of background watcher goroutines.
 	watcherCtx, watcherCancel := context.WithCancel(context.Background())
 
-	// Create server with initial handler
+	// Create server with initial handlergit
 	srv := &http.Server{
 		Addr: *listenStr,
 	}
@@ -140,7 +145,9 @@ func main() {
 
 			mainLogger.Debug("Configuration Changed")
 			currentPM.Shutdown()
-			mon.UpdateConfig(conf.Performance)
+			if mon != nil {
+				mon.UpdateConfig(conf.Performance)
+			}
 			newPM := proxy.New(conf)
 			newPM.SetVersion(date, commit, version)
 			newPM.SetPerfMonitor(mon)
@@ -197,7 +204,9 @@ func main() {
 				reloadProxyManager()
 			case syscall.SIGINT, syscall.SIGTERM:
 				mainLogger.Debugf("Received signal %v, shutting down...", sig)
-				mon.Stop()
+				if mon != nil {
+					mon.Stop()
+				}
 				watcherCancel()
 				ctx, cancel := context.WithTimeout(context.Background(), time.Second*5)
 				defer cancel()

proxy/config/config.go

@@ -223,11 +223,8 @@ func LoadConfigFromReader(r io.Reader) (Config, error) {
 	}
 
 	// Apply defaults for performance config when section is missing
-	if !config.Performance.Enable && config.Performance.Every == 0 && config.Performance.MaxAge == 0 && config.Performance.GC == 0 {
-		config.Performance.Enable = true
-		config.Performance.Every = 15 * time.Second
-		config.Performance.MaxAge = 1 * time.Hour
-		config.Performance.GC = 5 * time.Minute
+	if config.Performance.Every == 0 {
+		config.Performance.Every = 5 * time.Second
 	}
 	if err = config.Performance.Validate(); err != nil {
 		return Config{}, fmt.Errorf("performance: %w", err)

proxy/config/config_posix_test.go

@@ -231,10 +231,7 @@ groups:
 		MetricsMaxInMemory: 1000,
 		CaptureBuffer:      5,
 		Performance: PerformanceConfig{
-			Enable: true,
-			Every:  15 * time.Second,
-			MaxAge: 1 * time.Hour,
-			GC:     5 * time.Minute,
+			Every: 5 * time.Second,
 		},
 		Profiles: map[string][]string{
 			"test": {"model1", "model2"},

proxy/config/config_windows_test.go

@@ -220,10 +220,7 @@ groups:
 		MetricsMaxInMemory: 1000,
 		CaptureBuffer:      5,
 		Performance: PerformanceConfig{
-			Enable: true,
-			Every:  15 * time.Second,
-			MaxAge: 1 * time.Hour,
-			GC:     5 * time.Minute,
+			Every: 5 * time.Second,
 		},
 		Profiles: map[string][]string{
 			"test": {"model1", "model2"},

proxy/config/performance.go

@@ -7,19 +7,14 @@ import (
 
 // PerformanceConfig holds configuration for system performance monitoring
 type PerformanceConfig struct {
-	Enable bool          `yaml:"enable"`
-	Every  time.Duration `yaml:"every"`
-	MaxAge time.Duration `yaml:"maxAge"`
-	GC     time.Duration `yaml:"gc"`
+	Disabled bool          `yaml:"disabled"`
+	Every    time.Duration `yaml:"every"`
 }
 
 func (p *PerformanceConfig) UnmarshalYAML(unmarshal func(interface{}) error) error {
 	type rawPerformanceConfig PerformanceConfig
 	defaults := rawPerformanceConfig{
-		Enable: true,
-		Every:  15 * time.Second,
-		MaxAge: 1 * time.Hour,
-		GC:     5 * time.Minute,
+		Every: 5 * time.Second,
 	}
 
 	if err := unmarshal(&defaults); err != nil {
@@ -32,14 +27,8 @@ func (p *PerformanceConfig) UnmarshalYAML(unmarshal func(interface{}) error) err
 
 // Validate checks the PerformanceConfig values and returns an error if invalid
 func (p *PerformanceConfig) Validate() error {
-	if p.Every < time.Second {
-		return fmt.Errorf("every must be at least 1s, got %v", p.Every)
-	}
-	if p.MaxAge <= 0 {
-		return fmt.Errorf("maxAge must be greater than 0, got %v", p.MaxAge)
-	}
-	if p.GC <= 0 {
-		return fmt.Errorf("gc must be greater than 0, got %v", p.GC)
+	if p.Every < 5*time.Second {
+		return fmt.Errorf("every must be at least 5s, got %v", p.Every)
 	}
 	return nil
 }

proxy/config/performance_test.go

@@ -18,10 +18,8 @@ models:
 	assert.NoError(t, err)
 
 	// When performance section is missing, defaults should be applied
-	assert.True(t, config.Performance.Enable)
-	assert.Equal(t, 15*time.Second, config.Performance.Every)
-	assert.Equal(t, 1*time.Hour, config.Performance.MaxAge)
-	assert.Equal(t, 5*time.Minute, config.Performance.GC)
+	assert.False(t, config.Performance.Disabled)
+	assert.Equal(t, 5*time.Second, config.Performance.Every)
 }
 
 func TestPerformanceConfig_CustomValues(t *testing.T) {
@@ -29,8 +27,6 @@ func TestPerformanceConfig_CustomValues(t *testing.T) {
 performance:
   enable: true
   every: 30s
-  maxAge: 12h
-  gc: 10m
 models:
   model1:
     cmd: path/to/cmd --port ${PORT}
@@ -38,16 +34,14 @@ models:
 	config, err := LoadConfigFromReader(strings.NewReader(content))
 	assert.NoError(t, err)
 
-	assert.True(t, config.Performance.Enable)
+	assert.False(t, config.Performance.Disabled)
 	assert.Equal(t, 30*time.Second, config.Performance.Every)
-	assert.Equal(t, 12*time.Hour, config.Performance.MaxAge)
-	assert.Equal(t, 10*time.Minute, config.Performance.GC)
 }
 
 func TestPerformanceConfig_Disabled(t *testing.T) {
 	content := `
 performance:
-  enable: false
+  disabled: true
 models:
   model1:
     cmd: path/to/cmd --port ${PORT}
@@ -55,18 +49,15 @@ models:
 	config, err := LoadConfigFromReader(strings.NewReader(content))
 	assert.NoError(t, err)
 
-	assert.False(t, config.Performance.Enable)
+	assert.True(t, config.Performance.Disabled)
 	// Duration defaults should still apply
-	assert.Equal(t, 15*time.Second, config.Performance.Every)
-	assert.Equal(t, 1*time.Hour, config.Performance.MaxAge)
-	assert.Equal(t, 5*time.Minute, config.Performance.GC)
+	assert.Equal(t, 5*time.Second, config.Performance.Every)
 }
 
 func TestPerformanceConfig_PartialValues(t *testing.T) {
 	content := `
 performance:
   every: 10s
-  maxAge: 6h
 models:
   model1:
     cmd: path/to/cmd --port ${PORT}
@@ -75,58 +66,27 @@ models:
 	assert.NoError(t, err)
 
 	// enable should default to true
-	assert.True(t, config.Performance.Enable)
+	assert.False(t, config.Performance.Disabled)
 	assert.Equal(t, 10*time.Second, config.Performance.Every)
-	assert.Equal(t, 6*time.Hour, config.Performance.MaxAge)
-	// gc should use default
-	assert.Equal(t, 5*time.Minute, config.Performance.GC)
 }
 
 func TestPerformanceConfig_InvalidEvery(t *testing.T) {
 	content := `
 performance:
-  every: 500ms
+  every: 4s
 models:
   model1:
     cmd: path/to/cmd --port ${PORT}
 `
 	_, err := LoadConfigFromReader(strings.NewReader(content))
 	assert.Error(t, err)
-	assert.Contains(t, err.Error(), "every must be at least 1s")
-}
-
-func TestPerformanceConfig_InvalidMaxAge(t *testing.T) {
-	content := `
-performance:
-  maxAge: 0s
-models:
-  model1:
-    cmd: path/to/cmd --port ${PORT}
-`
-	_, err := LoadConfigFromReader(strings.NewReader(content))
-	assert.Error(t, err)
-	assert.Contains(t, err.Error(), "maxAge must be greater than 0")
-}
-
-func TestPerformanceConfig_InvalidGC(t *testing.T) {
-	content := `
-performance:
-  gc: 0s
-models:
-  model1:
-    cmd: path/to/cmd --port ${PORT}
-`
-	_, err := LoadConfigFromReader(strings.NewReader(content))
-	assert.Error(t, err)
-	assert.Contains(t, err.Error(), "gc must be greater than 0")
+	assert.Contains(t, err.Error(), "every must be at least 5s")
 }
 
 func TestPerformanceConfig_ComplexDurations(t *testing.T) {
 	content := `
 performance:
   every: 1m30s
-  maxAge: 2h10m
-  gc: 1m
 models:
   model1:
     cmd: path/to/cmd --port ${PORT}
@@ -135,6 +95,4 @@ models:
 	assert.NoError(t, err)
 
 	assert.Equal(t, 90*time.Second, config.Performance.Every)
-	assert.Equal(t, (2*time.Hour)+(10*time.Minute), config.Performance.MaxAge)
-	assert.Equal(t, 1*time.Minute, config.Performance.GC)
 }