fix(hermes): use WebUI for control surfaces

services/hermes/agent-deployment.yaml

@@ -227,9 +227,9 @@ spec:
           image: registry.bstein.dev/bstein/hermes-agent@sha256:15c5c538c0b58686af2e54e10bc870b23284789d485a609349df24ed3053622f
           imagePullPolicy: IfNotPresent
           command: [/opt/hermes/.venv/bin/hermes]
-          args: [gateway, run]
+          args: [gateway, run, --no-supervise]
           ports:
-            - {name: dashboard, containerPort: 9119, protocol: TCP}
+            - {name: api, containerPort: 8642, protocol: TCP}
           env:
             - {name: HERMES_HOME, value: /opt/data}
             - {name: HERMES_AUTH_FILE, value: /shared-auth/auth.json}
@@ -239,10 +239,12 @@ spec:
             - {name: HERDR_CONFIG_PATH, value: /opt/data/home/.config/herdr/config.toml}
             - {name: HERDR_SOCKET_PATH, value: /opt/data/herdr/herdr.sock}
             - {name: PATH, value: /opt/data/tools/bin:/opt/data/home/.local/bin:/opt/hermes/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin}
-            - {name: HERMES_DASHBOARD, value: "1"}
+            - {name: HERMES_DASHBOARD, value: "0"}
-            - {name: HERMES_DASHBOARD_HOST, value: 0.0.0.0}
-            - {name: HERMES_DASHBOARD_PORT, value: "9119"}
             - {name: HERMES_DASHBOARD_PUBLIC_URL, value: https://agent.hermes.bstein.dev}
+            - {name: API_SERVER_ENABLED, value: "true"}
+            - {name: API_SERVER_HOST, value: 0.0.0.0}
+            - {name: API_SERVER_PORT, value: "8642"}
+            - {name: API_SERVER_CORS_ORIGINS, value: https://agent.hermes.bstein.dev}
           volumeMounts:
             - {name: home, mountPath: /opt/data}
             - {name: provider-auth, mountPath: /shared-auth}
@@ -250,17 +252,17 @@ spec:
             - {name: auth-patch, mountPath: /opt/hermes/hermes_cli/auth.py, subPath: auth.py}
             - {name: coordinator, mountPath: /opt/data/home/.local/bin/herdr-dispatch, subPath: herdr_dispatch.py, readOnly: true}
           startupProbe:
-            httpGet: {path: /api/status, port: dashboard}
+            tcpSocket: {port: api}
             periodSeconds: 10
             timeoutSeconds: 5
             failureThreshold: 60
           readinessProbe:
-            httpGet: {path: /api/status, port: dashboard}
+            tcpSocket: {port: api}
             initialDelaySeconds: 30
             periodSeconds: 10
             timeoutSeconds: 5
           livenessProbe:
-            httpGet: {path: /api/status, port: dashboard}
+            tcpSocket: {port: api}
             initialDelaySeconds: 90
             periodSeconds: 30
             timeoutSeconds: 10
@@ -273,6 +275,63 @@ spec:
           resources:
             requests: {cpu: 250m, memory: 512Mi}
             limits: {cpu: "2", memory: 4Gi}
+        - name: webui
+          image: registry.bstein.dev/bstein/hermes-webui@sha256:a771858bd668d25e19c74864baea5425101c8cd5215d1ba3a312f3312ce6c5e1
+          imagePullPolicy: IfNotPresent
+          command: [/bin/sh, -ec]
+          args:
+            - |
+              api_key="$(sed -n 's/^API_SERVER_KEY=//p' /opt/data/.env | tail -n 1)"
+              test -n "${api_key}"
+              export API_SERVER_KEY="${api_key}"
+              export HERMES_WEBUI_GATEWAY_API_KEY="${api_key}"
+              exec /opt/hermes/.venv/bin/python /opt/hermes-webui/server.py
+          ports:
+            - {name: dashboard, containerPort: 8787, protocol: TCP}
+          env:
+            - {name: HERMES_HOME, value: /opt/data}
+            - {name: HERMES_AUTH_FILE, value: /shared-auth/auth.json}
+            - {name: HOME, value: /opt/data/home}
+            - {name: HERMES_WEBUI_AGENT_DIR, value: /opt/hermes}
+            - {name: HERMES_WEBUI_HOST, value: 0.0.0.0}
+            - {name: HERMES_WEBUI_PORT, value: "8787"}
+            - {name: HERMES_WEBUI_STATE_DIR, value: /opt/data/webui}
+            - {name: HERMES_WEBUI_DEFAULT_WORKSPACE, value: /opt/data/workspace}
+            - {name: HERMES_WEBUI_CHAT_BACKEND, value: gateway}
+            - {name: HERMES_WEBUI_GATEWAY_BASE_URL, value: http://127.0.0.1:8642}
+            - {name: HERMES_WEBUI_GATEWAY_USE_RUNS_API, value: "true"}
+            - {name: HERMES_WEBUI_SKIP_ONBOARDING, value: "1"}
+            - {name: HERMES_WEBUI_SECURE, value: "1"}
+            - {name: HERMES_WEBUI_ALLOWED_ORIGINS, value: https://agent.hermes.bstein.dev}
+            - {name: HERMES_WEBUI_TRUST_FORWARDED_HOST, value: "1"}
+            - {name: HERMES_WEBUI_TRUST_FORWARDED_PROTO, value: "1"}
+          volumeMounts:
+            - {name: home, mountPath: /opt/data}
+            - {name: provider-auth, mountPath: /shared-auth, readOnly: true}
+            - {name: tmp, mountPath: /tmp}
+          readinessProbe:
+            httpGet: {path: /health, port: dashboard}
+            initialDelaySeconds: 10
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            httpGet: {path: /health, port: dashboard}
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 10
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: [ALL]
+            readOnlyRootFilesystem: true
+            runAsNonRoot: true
+            runAsUser: 10000
+            runAsGroup: 10000
+            seccompProfile:
+              type: RuntimeDefault
+          resources:
+            requests: {cpu: 50m, memory: 128Mi}
+            limits: {cpu: 750m, memory: 1Gi}
         - name: herdr-server
           image: registry.bstein.dev/bstein/hermes-agent@sha256:15c5c538c0b58686af2e54e10bc870b23284789d485a609349df24ed3053622f
           imagePullPolicy: IfNotPresent
@@ -351,3 +410,6 @@ spec:
             defaultMode: 0555
         - name: auth-patch
           emptyDir: {}
+        - name: tmp
+          emptyDir:
+            sizeLimit: 256Mi

services/hermes/deployment.yaml

@@ -242,13 +242,11 @@ spec:
           args:
             - gateway
             - run
+            - --no-supervise
           ports:
             - name: api
               containerPort: 8642
               protocol: TCP
-            - name: dashboard
-              containerPort: 9119
-              protocol: TCP
           env:
             - name: HERMES_HOME
               value: /opt/data
@@ -259,11 +257,7 @@ spec:
             - name: PATH
               value: /opt/data/home/.local/bin:/opt/hermes/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
             - name: HERMES_DASHBOARD
-              value: "1"
+              value: "0"
-            - name: HERMES_DASHBOARD_HOST
-              value: 0.0.0.0
-            - name: HERMES_DASHBOARD_PORT
-              value: "9119"
             - name: HERMES_DASHBOARD_PUBLIC_URL
               value: https://triage.hermes.bstein.dev
             - name: API_SERVER_ENABLED
@@ -325,23 +319,20 @@ spec:
               mountPath: /opt/data/workspace/skills/tune-atlas-alerts
               readOnly: true
           startupProbe:
-            httpGet:
+            tcpSocket:
-              path: /api/status
+              port: api
-              port: dashboard
             periodSeconds: 10
             timeoutSeconds: 5
             failureThreshold: 60
           readinessProbe:
-            httpGet:
+            tcpSocket:
-              path: /api/status
+              port: api
-              port: dashboard
             initialDelaySeconds: 30
             periodSeconds: 10
             timeoutSeconds: 5
           livenessProbe:
-            httpGet:
+            tcpSocket:
-              path: /api/status
+              port: api
-              port: dashboard
             initialDelaySeconds: 90
             periodSeconds: 30
             timeoutSeconds: 10
@@ -352,6 +343,63 @@ spec:
             limits:
               cpu: "2"
               memory: 4Gi
+        - name: webui
+          image: registry.bstein.dev/bstein/hermes-webui@sha256:a771858bd668d25e19c74864baea5425101c8cd5215d1ba3a312f3312ce6c5e1
+          imagePullPolicy: IfNotPresent
+          command: [/bin/sh, -ec]
+          args:
+            - |
+              api_key="$(sed -n 's/^API_SERVER_KEY=//p' /opt/data/.env | tail -n 1)"
+              test -n "${api_key}"
+              export API_SERVER_KEY="${api_key}"
+              export HERMES_WEBUI_GATEWAY_API_KEY="${api_key}"
+              exec /opt/hermes/.venv/bin/python /opt/hermes-webui/server.py
+          ports:
+            - {name: dashboard, containerPort: 8787, protocol: TCP}
+          env:
+            - {name: HERMES_HOME, value: /opt/data}
+            - {name: HERMES_AUTH_FILE, value: /shared-auth/auth.json}
+            - {name: HOME, value: /opt/data/home}
+            - {name: HERMES_WEBUI_AGENT_DIR, value: /opt/hermes}
+            - {name: HERMES_WEBUI_HOST, value: 0.0.0.0}
+            - {name: HERMES_WEBUI_PORT, value: "8787"}
+            - {name: HERMES_WEBUI_STATE_DIR, value: /opt/data/webui}
+            - {name: HERMES_WEBUI_DEFAULT_WORKSPACE, value: /opt/data/workspace}
+            - {name: HERMES_WEBUI_CHAT_BACKEND, value: gateway}
+            - {name: HERMES_WEBUI_GATEWAY_BASE_URL, value: http://127.0.0.1:8642}
+            - {name: HERMES_WEBUI_GATEWAY_USE_RUNS_API, value: "true"}
+            - {name: HERMES_WEBUI_SKIP_ONBOARDING, value: "1"}
+            - {name: HERMES_WEBUI_SECURE, value: "1"}
+            - {name: HERMES_WEBUI_ALLOWED_ORIGINS, value: https://triage.hermes.bstein.dev}
+            - {name: HERMES_WEBUI_TRUST_FORWARDED_HOST, value: "1"}
+            - {name: HERMES_WEBUI_TRUST_FORWARDED_PROTO, value: "1"}
+          volumeMounts:
+            - {name: home, mountPath: /opt/data}
+            - {name: provider-auth, mountPath: /shared-auth, readOnly: true}
+            - {name: tmp, mountPath: /tmp}
+          readinessProbe:
+            httpGet: {path: /health, port: dashboard}
+            initialDelaySeconds: 10
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            httpGet: {path: /health, port: dashboard}
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 10
+          securityContext:
+            allowPrivilegeEscalation: false
+            capabilities:
+              drop: [ALL]
+            readOnlyRootFilesystem: true
+            runAsNonRoot: true
+            runAsUser: 10000
+            runAsGroup: 10000
+            seccompProfile:
+              type: RuntimeDefault
+          resources:
+            requests: {cpu: 50m, memory: 128Mi}
+            limits: {cpu: 750m, memory: 1Gi}
       volumes:
         - name: home
           persistentVolumeClaim:
@@ -373,6 +421,9 @@ spec:
             defaultMode: 0555
         - name: auth-patch
           emptyDir: {}
+        - name: tmp
+          emptyDir:
+            sizeLimit: 256Mi
         - name: triage-skill
           configMap:
             name: hermes-triage-skill