hermes(agent): retry dashboard ticket acquisition

2026-08-10 23:57:32 -03:00 · 2026-08-10 23:57:32 -03:00 · 987c6e4cee
commit 987c6e4cee
parent 0bec6edba3
3 changed files with 71 additions and 12 deletions
--- a/dockerfiles/Dockerfile.hermes-agent
+++ b/dockerfiles/Dockerfile.hermes-agent
@ -27,7 +27,13 @@ const socketBefore = [
  '      const ws = new WebSocket(url);',
 ].join("\n");
 const socketAfter = [
-  '      const url = await api.buildWsUrl("/api/pty", params);',
+  '      let url: string;',
+  '      try {',
+  '        url = await api.buildWsUrl("/api/pty", params);',
+  '      } catch {',
+  '        if (!unmounting) scheduleReconnect(1006);',
+  '        return;',
+  '      }',
  '      if (unmounting) return;',
  '      const ws = new WebSocket(url);',
 ].join("\n");
@ -70,6 +76,32 @@ const socketStateAfter = [
  '    let ws: WebSocket | null = null;',
  '    let reconnectTimer: ReturnType<typeof setTimeout> | null = null;',
 ].join("\n");
+const ticketBefore = [
+  '      const url = await buildWsUrl("/api/events", { channel });',
+  '      if (unmounting) {',
+  '        return;',
+  '      }',
+  '      ws = new WebSocket(url);',
+].join("\n");
+const ticketAfter = [
+  '      let url: string;',
+  '      try {',
+  '        url = await buildWsUrl("/api/events", { channel });',
+  '      } catch {',
+  '        if (unmounting) return;',
+  '        const attempt = Math.min(eventsRetryAttempt.current + 1, 6);',
+  '        eventsRetryAttempt.current = attempt;',
+  '        const delay = Math.min(500 * 2 ** (attempt - 1), 5000);',
+  '        setError("events feed disconnected — tool calls may not appear; reconnecting…");',
+  '        reconnectTimer = setTimeout(() => {',
+  '          reconnectTimer = null;',
+  '          if (!unmounting) setVersion((v) => v + 1);',
+  '        }, delay);',
+  '        return;',
+  '      }',
+  '      if (unmounting) return;',
+  '      ws = new WebSocket(url);',
+].join("\n");
 const handlersBefore = [
  '      ws.addEventListener("error", () => surface(DISCONNECTED));',
  '',
@ -121,6 +153,7 @@ const cleanupAfter = [
 for (const [before, after, label] of [
  [retryStateBefore, retryStateAfter, "retry state"],
  [socketStateBefore, socketStateAfter, "socket state"],
+  [ticketBefore, ticketAfter, "ticket retry"],
  [handlersBefore, handlersAfter, "socket handlers"],
  [cleanupBefore, cleanupAfter, "socket cleanup"],
 ]) {
--- a/services/hermes/agent-deployment.yaml
+++ b/services/hermes/agent-deployment.yaml
@ -170,7 +170,7 @@ spec:
            requests: {cpu: 25m, memory: 32Mi}
            limits: {cpu: 100m, memory: 64Mi}
        - name: install-agent-tools
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - sh
@ -218,7 +218,7 @@ spec:
            requests: {cpu: 100m, memory: 256Mi}
            limits: {cpu: "1", memory: 1Gi}
        - name: patch-auth
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - /opt/hermes/.venv/bin/python
@ -241,7 +241,7 @@ spec:
            requests: {cpu: 25m, memory: 64Mi}
            limits: {cpu: 100m, memory: 128Mi}
        - name: patch-tui-gateway
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - /opt/hermes/.venv/bin/python
@ -264,7 +264,7 @@ spec:
            requests: {cpu: 25m, memory: 64Mi}
            limits: {cpu: 100m, memory: 128Mi}
        - name: patch-codex-runtime
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - /opt/hermes/.venv/bin/python
@ -295,7 +295,7 @@ spec:
            requests: {cpu: 25m, memory: 64Mi}
            limits: {cpu: 100m, memory: 128Mi}
        - name: bootstrap-coordinator
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - /opt/hermes/.venv/bin/python
@ -305,7 +305,10 @@ spec:
            - {name: HERMES_HOME, value: /opt/data}
            - {name: HERMES_AUTH_FILE, value: /shared-auth/auth.json}
            - {name: HOME, value: /opt/data/home}
+            - {name: CODEX_HOME, value: /opt/data/home/.codex}
+            - {name: CLAUDE_CONFIG_DIR, value: /opt/data/home/.claude}
            - {name: PYTHONPATH, value: /opt/hermes}
+            - {name: PATH, value: /opt/coordinator:/opt/data/tools/bin:/opt/hermes/.venv/bin:/usr/local/bin:/usr/bin:/bin}
          securityContext:
            allowPrivilegeEscalation: false
            runAsUser: 10000
@ -321,7 +324,7 @@ spec:
            requests: {cpu: 50m, memory: 128Mi}
            limits: {cpu: 500m, memory: 512Mi}
        - name: configure-agent-clients
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - sh
@ -359,7 +362,7 @@ spec:
            requests: {cpu: 25m, memory: 32Mi}
            limits: {cpu: 250m, memory: 128Mi}
        - name: prepare-ttyd-index
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command:
            - /opt/hermes/.venv/bin/python
@ -381,7 +384,7 @@ spec:
            limits: {cpu: 250m, memory: 128Mi}
      containers:
        - name: hermes
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command: [/init, /opt/hermes/docker/main-wrapper.sh]
          args: [gateway, run]
@ -512,7 +515,7 @@ spec:
            - {name: allowlist, mountPath: /etc/oauth2-proxy, readOnly: true}
            - {name: oauth-tmp, mountPath: /tmp}
        - name: terminal
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command: [/bin/sh, -ec]
          args:
@ -604,7 +607,7 @@ spec:
            requests: {cpu: 25m, memory: 64Mi}
            limits: {cpu: 500m, memory: 512Mi}
        - name: cli-lane-runner
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command: [/bin/sh, -ec]
          args:
@ -645,14 +648,17 @@ spec:
            requests: {cpu: 100m, memory: 256Mi}
            limits: {cpu: "3", memory: 6Gi}
        - name: model-steward
-          image: registry.bstein.dev/bstein/hermes-agent@sha256:0577eb22d28b1ffc649405d31919b2527d9e591ccaa4a1e168dd6d6392a5a43f
+          image: registry.bstein.dev/bstein/hermes-agent@sha256:cf07be056feea8e4f2d5512899f990732f3d26b4915257de8f90105d96cec67d
          imagePullPolicy: IfNotPresent
          command: [/opt/hermes/.venv/bin/python, /opt/coordinator/hermes_coordinator.py, --loop, --interval, "3600"]
          env:
            - {name: HERMES_HOME, value: /opt/data}
            - {name: HERMES_AUTH_FILE, value: /shared-auth/auth.json}
            - {name: HOME, value: /opt/data/home}
+            - {name: CODEX_HOME, value: /opt/data/home/.codex}
+            - {name: CLAUDE_CONFIG_DIR, value: /opt/data/home/.claude}
            - {name: PYTHONPATH, value: /opt/hermes}
+            - {name: PATH, value: /opt/coordinator:/opt/data/tools/bin:/opt/hermes/.venv/bin:/usr/local/bin:/usr/bin:/bin}
          securityContext:
            allowPrivilegeEscalation: false
            runAsUser: 10000
--- a/testing/tests/test_hermes_cli_lanes.py
+++ b/testing/tests/test_hermes_cli_lanes.py
@ -723,6 +723,8 @@ def test_agent_dashboard_reconnects_all_transient_websockets():
    assert "eventsRetryAttempt.current" in dockerfile
    assert "if (!unmounting) setVersion((v) => v + 1);" in dockerfile
    assert "events feed rejected (${ev.code}) — reload the page" in dockerfile
+    assert 'url = await api.buildWsUrl("/api/pty", params);' in dockerfile
+    assert 'url = await buildWsUrl("/api/events", { channel });' in dockerfile


 def test_agent_refreshes_routes_after_restoring_cli_logins():
@ -740,6 +742,24 @@ def test_agent_refreshes_routes_after_restoring_cli_logins():
    env = {item["name"]: item["value"] for item in configure["env"]}
    assert env["HERMES_AUTH_FILE"] == "/shared-auth/auth.json"
    assert env["PYTHONPATH"] == "/opt/hermes"
+    for name in ("bootstrap-coordinator", "configure-agent-clients"):
+        route_env = {
+            item["name"]: item["value"]
+            for item in init_containers[name]["env"]
+        }
+        assert route_env["CODEX_HOME"] == "/opt/data/home/.codex"
+        assert "/opt/data/tools/bin" in route_env["PATH"]
+
+    containers = {
+        item["name"]: item
+        for item in deployment["spec"]["template"]["spec"]["containers"]
+    }
+    steward_env = {
+        item["name"]: item["value"]
+        for item in containers["model-steward"]["env"]
+    }
+    assert steward_env["CODEX_HOME"] == "/opt/data/home/.codex"
+    assert "/opt/data/tools/bin" in steward_env["PATH"]


 def test_flux_health_checks_follow_the_owner_oauth_sidecar():