feat(server): adds minimum keepalive and headroom, chore: updates packages to fix vulnerabilities

src/Ngino.Client/Ngino.Client.csproj

@@ -5,7 +5,7 @@
   </ItemGroup>
 
   <ItemGroup>
-    <PackageReference Include="Microsoft.Extensions.Hosting.WindowsServices" Version="8.0.1" />
+    <PackageReference Include="Microsoft.Extensions.Hosting.WindowsServices" Version="10.0.10" />
   </ItemGroup>
 
   <PropertyGroup>

src/Ngino.Client/Properties/AssemblyInfo.cs

@@ -1,3 +1,4 @@
 using System.Runtime.CompilerServices;
 
 [assembly: InternalsVisibleTo("Ngino.Client.Tests")]
+[assembly: InternalsVisibleTo("ReverseLlama.Client.Tests")]

src/Ngino.Server/AdminEndpoints.cs

@@ -369,7 +369,14 @@ internal static class AdminEndpoints
 
             try
             {
-                var member = store.AddGroupClient(id, request.ClientId, request.Model, request.ClientPattern);
+                var member = store.AddGroupClient(
+                    id,
+                    request.ClientId,
+                    request.Model,
+                    request.ClientPattern,
+                    request.KeepaliveInstancesToKeepAlive,
+                    request.KeepaliveMaxParallelismPerClient,
+                    request.KeepaliveParallelismHeadroom);
                 return Results.Json(member);
             }
             catch (ArgumentException exception)
@@ -889,7 +896,10 @@ internal sealed record UpdateGroupRequest(string Name);
 internal sealed record AddGroupClientRequest(
     string? ClientId,
     string? Model,
-    string? ClientPattern);
+    string? ClientPattern,
+    int? KeepaliveInstancesToKeepAlive,
+    int? KeepaliveMaxParallelismPerClient,
+    int? KeepaliveParallelismHeadroom);
 
 internal sealed record SetUserKeyGroupsRequest(IReadOnlyList<string>? GroupIds);
 

src/Ngino.Server/ManagementStore.cs

@@ -824,7 +824,10 @@ internal sealed class ManagementStore
             using var connection = OpenConnection();
             using var command = connection.CreateCommand();
             command.CommandText = """
-                SELECT id, group_id, client_id, model, client_pattern
+                SELECT id, group_id, client_id, model, client_pattern,
+                       keepalive_instances_to_keep_alive,
+                       keepalive_max_parallelism_per_client,
+                       keepalive_parallelism_headroom
                 FROM group_members
                 WHERE group_id = $group_id
                 ORDER BY client_id, model, client_pattern
@@ -840,14 +843,22 @@ internal sealed class ManagementStore
                     reader.GetString(1),
                     reader.IsDBNull(2) ? null : reader.GetString(2),
                     reader.IsDBNull(3) ? null : reader.GetString(3),
-                    reader.IsDBNull(4) ? null : reader.GetString(4)));
+                    reader.IsDBNull(4) ? null : reader.GetString(4),
+                    ReadKeepalivePolicy(reader, 5, 6, 7)));
             }
 
             return result;
         }
     }
 
-    public GroupClientInfo AddGroupClient(string groupId, string? clientId, string? model, string? clientPattern)
+    public GroupClientInfo AddGroupClient(
+        string groupId,
+        string? clientId,
+        string? model,
+        string? clientPattern,
+        int? keepaliveInstancesToKeepAlive,
+        int? keepaliveMaxParallelismPerClient,
+        int? keepaliveParallelismHeadroom)
     {
         EnsureAvailable();
 
@@ -873,24 +884,46 @@ internal sealed class ManagementStore
             }
         }
 
+        var policy = NormalizeKeepalivePolicy(
+            keepaliveInstancesToKeepAlive,
+            keepaliveMaxParallelismPerClient,
+            keepaliveParallelismHeadroom);
+
         lock (_lock)
         {
             using var connection = OpenConnection();
             using var command = connection.CreateCommand();
             command.CommandText = """
-                INSERT INTO group_members (group_id, client_id, model, client_pattern)
-                VALUES ($group_id, $client_id, $model, $client_pattern)
+                INSERT INTO group_members (
+                    group_id,
+                    client_id,
+                    model,
+                    client_pattern,
+                    keepalive_instances_to_keep_alive,
+                    keepalive_max_parallelism_per_client,
+                    keepalive_parallelism_headroom)
+                VALUES (
+                    $group_id,
+                    $client_id,
+                    $model,
+                    $client_pattern,
+                    $keepalive_instances_to_keep_alive,
+                    $keepalive_max_parallelism_per_client,
+                    $keepalive_parallelism_headroom)
                 """;
             command.Parameters.AddWithValue("$group_id", groupId);
             command.Parameters.AddWithValue("$client_id", string.IsNullOrWhiteSpace(clientId) ? DBNull.Value : clientId);
             command.Parameters.AddWithValue("$model", string.IsNullOrWhiteSpace(model) ? DBNull.Value : model);
             command.Parameters.AddWithValue("$client_pattern", string.IsNullOrWhiteSpace(clientPattern) ? DBNull.Value : clientPattern);
+            command.Parameters.AddWithValue("$keepalive_instances_to_keep_alive", policy.InstancesToKeepAlive);
+            command.Parameters.AddWithValue("$keepalive_max_parallelism_per_client", policy.MaxParallelismPerClient);
+            command.Parameters.AddWithValue("$keepalive_parallelism_headroom", policy.ParallelismHeadroom);
             command.ExecuteNonQuery();
 
             using var idCommand = connection.CreateCommand();
             idCommand.CommandText = "SELECT last_insert_rowid()";
             var insertedId = (long)idCommand.ExecuteScalar()!;
-            return new GroupClientInfo(insertedId, groupId, clientId, model, clientPattern);
+            return new GroupClientInfo(insertedId, groupId, clientId, model, clientPattern, policy);
         }
     }
 
@@ -1987,7 +2020,10 @@ internal sealed class ManagementStore
                     group_id TEXT NOT NULL REFERENCES groups(id) ON DELETE CASCADE,
                     client_id TEXT NULL,
                     model TEXT NULL,
-                    client_pattern TEXT NULL
+                    client_pattern TEXT NULL,
+                    keepalive_instances_to_keep_alive INTEGER NOT NULL DEFAULT 1,
+                    keepalive_max_parallelism_per_client INTEGER NOT NULL DEFAULT 1,
+                    keepalive_parallelism_headroom INTEGER NOT NULL DEFAULT 1
                 );
 
                 CREATE INDEX IF NOT EXISTS idx_group_members_group_id
@@ -2010,6 +2046,29 @@ internal sealed class ManagementStore
                 """;
             command.ExecuteNonQuery();
 
+        using (var migrate = connection.CreateCommand())
+        {
+            migrate.CommandText = """
+                SELECT COUNT(*) FROM pragma_table_info('group_members') WHERE name = 'keepalive_instances_to_keep_alive'
+                """;
+            var hasKeepaliveInstancesColumn = (long)migrate.ExecuteScalar()! > 0;
+
+            if (!hasKeepaliveInstancesColumn)
+            {
+                using var addKeepaliveInstances = connection.CreateCommand();
+                addKeepaliveInstances.CommandText = "ALTER TABLE group_members ADD COLUMN keepalive_instances_to_keep_alive INTEGER NOT NULL DEFAULT 1";
+                addKeepaliveInstances.ExecuteNonQuery();
+
+                using var addKeepaliveMaxParallelism = connection.CreateCommand();
+                addKeepaliveMaxParallelism.CommandText = "ALTER TABLE group_members ADD COLUMN keepalive_max_parallelism_per_client INTEGER NOT NULL DEFAULT 1";
+                addKeepaliveMaxParallelism.ExecuteNonQuery();
+
+                using var addKeepaliveHeadroom = connection.CreateCommand();
+                addKeepaliveHeadroom.CommandText = "ALTER TABLE group_members ADD COLUMN keepalive_parallelism_headroom INTEGER NOT NULL DEFAULT 1";
+                addKeepaliveHeadroom.ExecuteNonQuery();
+            }
+        }
+
         using (var migrate = connection.CreateCommand())
         {
             migrate.CommandText = """
@@ -2231,6 +2290,30 @@ internal sealed class ManagementStore
         return connection;
     }
 
+    private static GroupClientKeepalivePolicy? ReadKeepalivePolicy(SqliteDataReader reader, int instancesOrdinal, int maxParallelismOrdinal, int headroomOrdinal)
+    {
+        if (reader.IsDBNull(instancesOrdinal) || reader.IsDBNull(maxParallelismOrdinal) || reader.IsDBNull(headroomOrdinal))
+        {
+            return null;
+        }
+
+        return NormalizeKeepalivePolicy(
+            reader.GetInt32(instancesOrdinal),
+            reader.GetInt32(maxParallelismOrdinal),
+            reader.GetInt32(headroomOrdinal));
+    }
+
+    private static GroupClientKeepalivePolicy NormalizeKeepalivePolicy(
+        int? instancesToKeepAlive,
+        int? maxParallelismPerClient,
+        int? parallelismHeadroom)
+    {
+        return new GroupClientKeepalivePolicy(
+            Math.Max(1, instancesToKeepAlive ?? GroupClientKeepalivePolicy.Default.InstancesToKeepAlive),
+            Math.Max(1, maxParallelismPerClient ?? GroupClientKeepalivePolicy.Default.MaxParallelismPerClient),
+            Math.Max(1, parallelismHeadroom ?? GroupClientKeepalivePolicy.Default.ParallelismHeadroom));
+    }
+
     private void EnsureAvailable()
     {
         if (!_isAvailable)
@@ -2420,7 +2503,16 @@ internal sealed record GroupClientInfo(
     string GroupId,
     string? ClientId,
     string? Model,
-    string? ClientPattern);
+    string? ClientPattern,
+    GroupClientKeepalivePolicy? KeepalivePolicy = null);
+
+internal sealed record GroupClientKeepalivePolicy(
+    int InstancesToKeepAlive,
+    int MaxParallelismPerClient,
+    int ParallelismHeadroom)
+{
+    public static GroupClientKeepalivePolicy Default { get; } = new(1, 1, 1);
+}
 
 internal sealed record UserKeyGroupInfo(
     string UserKeyId,

src/Ngino.Server/Ngino.Server.csproj

@@ -3,14 +3,20 @@
   <ItemGroup>
     <PackageReference Include="ElmahCore" Version="2.1.2" />
     <PackageReference Include="ElmahCore.MySql" Version="2.1.2" />
-    <PackageReference Include="Microsoft.AspNetCore.Authentication.OpenIdConnect" Version="8.0.28" />
-    <PackageReference Include="Microsoft.AspNetCore.Identity.EntityFrameworkCore" Version="8.0.0" />
-    <PackageReference Include="Microsoft.Data.Sqlite" Version="8.0.28" />
-    <PackageReference Include="Microsoft.EntityFrameworkCore.Sqlite" Version="8.0.0" />
-    <PackageReference Include="SQLitePCLRaw.bundle_e_sqlite3" Version="3.0.3" />
-    <PackageReference Include="System.Text.Encodings.Web" Version="8.0.0" />
-    <PackageReference Include="System.Text.Json" Version="8.0.5" />
+    <PackageReference Include="Microsoft.AspNetCore.Authentication.OpenIdConnect" Version="10.0.10" />
+    <PackageReference Include="Microsoft.AspNetCore.Identity.EntityFrameworkCore" Version="10.0.10" />
+    <PackageReference Include="Microsoft.Data.Sqlite" Version="10.0.10" />
+    <PackageReference Include="Microsoft.EntityFrameworkCore.Sqlite" Version="10.0.10" />
+    <PackageReference Include="SSH.NET" Version="2024.1.0" />
+    <PackageReference Include="SQLitePCLRaw.bundle_e_sqlite3" Version="3.0.5" />
+    <PackageReference Include="System.Data.SqlClient" Version="4.9.1" />
+    <PackageReference Include="System.Drawing.Common" Version="9.0.0" />
+    <PackageReference Include="System.Text.Encodings.Web" Version="10.0.10" />
+    <PackageReference Include="System.Text.Json" Version="10.0.10" />
     <ProjectReference Include="..\Ngino.Protocol\Ngino.Protocol.csproj" />
+    <AssemblyAttribute Include="System.Runtime.CompilerServices.InternalsVisibleTo">
+      <_Parameter1>Ngino.Client.Tests</_Parameter1>
+    </AssemblyAttribute>
   </ItemGroup>
 
   <PropertyGroup>

src/Ngino.Server/Properties/AssemblyInfo.cs

@@ -0,0 +1,3 @@
+using System.Runtime.CompilerServices;
+
+[assembly: InternalsVisibleTo("ReverseLlama.Client.Tests")]

src/Ngino.Server/wwwroot/admin/app.js

@@ -228,6 +228,9 @@ content.addEventListener("submit", async (event) => {
       if (data.clientPattern && data.clientPattern.trim()) {
         body.clientPattern = data.clientPattern.trim();
       }
+      body.keepaliveInstancesToKeepAlive = parseInt(data.keepaliveInstancesToKeepAlive, 10) || 1;
+      body.keepaliveMaxParallelismPerClient = parseInt(data.keepaliveMaxParallelismPerClient, 10) || 1;
+      body.keepaliveParallelismHeadroom = parseInt(data.keepaliveParallelismHeadroom, 10) || 1;
 
       if (!body.clientId && !body.clientPattern) {
         throw new Error("Either Client ID or Client pattern is required.");
@@ -981,9 +984,21 @@ function renderGroupDetail() {
             <label for="addClientPattern">Client pattern (regex, optional)</label>
             <input class="input" id="addClientPattern" name="clientPattern" placeholder="GPU_[0-9]*">
           </div>
+          <div class="field">
+            <label for="addClientKeepaliveInstances">Keepalive instances</label>
+            <input class="input" id="addClientKeepaliveInstances" name="keepaliveInstancesToKeepAlive" type="number" min="1" step="1" value="1">
+          </div>
+          <div class="field">
+            <label for="addClientKeepaliveMaxParallelism">Max parallelism per client</label>
+            <input class="input" id="addClientKeepaliveMaxParallelism" name="keepaliveMaxParallelismPerClient" type="number" min="1" step="1" value="1">
+          </div>
+          <div class="field">
+            <label for="addClientKeepaliveHeadroom">Parallelism headroom</label>
+            <input class="input" id="addClientKeepaliveHeadroom" name="keepaliveParallelismHeadroom" type="number" min="1" step="1" value="1">
+          </div>
           <button class="button" type="submit">Add</button>
         </form>
-        <div class="cell-sub" style="margin-top:8px">Provide either a Client ID (for explicit client access) or a Client pattern (for regex-based matching). Model is optional to restrict to a specific model.</div>
+        <div class="cell-sub" style="margin-top:8px">Provide either a Client ID (for explicit client access) or a Client pattern (for regex-based matching). Model is optional to restrict to a specific model. Keepalive defaults to 1 instance, 1 parallelism, and 1 headroom.</div>
       </div>
     </div>
     <div class="panel">
@@ -1096,6 +1111,17 @@ function renderGroupDetail() {
   `);
 }
 
+function formatKeepalivePolicy(policy) {
+  if (!policy) {
+    return `<div class="cell-sub">Default (1 / 1 / 1)</div>`;
+  }
+
+  return `
+    <div class="cell-main">${policy.instancesToKeepAlive} instance${policy.instancesToKeepAlive === 1 ? "" : "s"}</div>
+    <div class="cell-sub">Parallelism ${policy.maxParallelismPerClient} · headroom ${policy.parallelismHeadroom}</div>
+  `;
+}
+
 function groupClientsTable(clients, groupId) {
   const rows = clients.map((client) => `
     <tr>
@@ -1117,6 +1143,7 @@ function groupClientsTable(clients, groupId) {
           : `<div class="cell-sub">-</div>`
         }
       </td>
+      <td>${formatKeepalivePolicy(client.keepalivePolicy)}</td>
       <td>
         <button class="button danger" data-action="remove-client" data-group-id="${escapeAttr(groupId)}" data-member-id="${client.id}">Remove</button>
       </td>
@@ -1130,6 +1157,7 @@ function groupClientsTable(clients, groupId) {
           <th>Client ID</th>
           <th>Model</th>
           <th>Client pattern</th>
+          <th>Keepalive policy</th>
           <th></th>
         </tr>
       </thead>

tests/Ngino.Client.Tests/GroupKeepalivePolicyTests.cs

@@ -0,0 +1,18 @@
+using Ngino.Server;
+using Xunit;
+
+namespace ReverseLlama.Client.Tests;
+
+public class GroupKeepalivePolicyTests
+{
+    [Fact]
+    public void KeepalivePolicy_IsRoundTripped_ThroughGroupClientInfo()
+    {
+        var policy = new GroupClientKeepalivePolicy(2, 3, 4);
+        var info = new GroupClientInfo(1, "group-1", "client-1", "model-a", "pattern", policy);
+
+        Assert.Equal(2, info.KeepalivePolicy?.InstancesToKeepAlive);
+        Assert.Equal(3, info.KeepalivePolicy?.MaxParallelismPerClient);
+        Assert.Equal(4, info.KeepalivePolicy?.ParallelismHeadroom);
+    }
+}

tests/Ngino.Client.Tests/Ngino.Client.Tests.csproj

@@ -9,9 +9,9 @@
   </PropertyGroup>
 
   <ItemGroup>
-    <PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.13.0" />
-    <PackageReference Include="xunit" Version="2.9.2" />
-    <PackageReference Include="xunit.runner.visualstudio" Version="2.8.2">
+    <PackageReference Include="Microsoft.NET.Test.Sdk" Version="18.8.1" />
+    <PackageReference Include="xunit" Version="2.9.3" />
+    <PackageReference Include="xunit.runner.visualstudio" Version="3.1.5">
       <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
       <PrivateAssets>all</PrivateAssets>
     </PackageReference>
@@ -19,6 +19,7 @@
 
   <ItemGroup>
     <ProjectReference Include="..\..\src\Ngino.Client\Ngino.Client.csproj" />
+    <ProjectReference Include="..\..\src\Ngino.Server\Ngino.Server.csproj" />
   </ItemGroup>
 
 </Project>