MoShiro/backend/src/Monitor.Backend.Application/Agents/MonitorFleetHooks.cs

using System.Text.Json;
using Fleet.Core;
using Fleet.Rpc;
using Fleet.Server;
using Monitor.Protocol;

namespace Monitor.Backend.Application;

/// <summary>Monitor 在 Fleet 之上的接入点：注册之后补全公开地址与端口。</summary>
public sealed class MonitorEnrollmentHook : IEnrollmentHook
{
    private readonly IAgentStore _agents;
    private readonly AgentService _service;

    public MonitorEnrollmentHook(IAgentStore agents, AgentService service)
    {
        _agents = agents;
        _service = service;
    }

    public async ValueTask OnEnrolledAsync(NodeRecord node, EnrollmentContext context, CancellationToken ct)
    {
        AgentEnrollMeta meta = new();
        if (context.ProductMeta is { } raw)
        {
            try
            {
                meta = raw.Deserialize(MonitorJsonContext.Default.AgentEnrollMeta) ?? meta;
            }
            catch (JsonException)
            {
                // 元数据格式不对就用默认值，节点之后可以通过 hello 补全
            }
        }

        // public_host 未提供时，取注册请求的来源 IP
        string host = !string.IsNullOrWhiteSpace(meta.PublicHost) ? meta.PublicHost : context.RemoteIp ?? "";
        await _agents.ApplyEnrollmentAsync(node.Id, new AgentEnrollment(host, meta.ProbeTcpPort, meta.ProbeUdpPorts, meta.Kinds), ct);
        if (node.Status == NodeStatus.Active)
        {
            await _service.AfterAgentSetChangedAsync(ct);
        }
    }
}

/// <summary>节点上线、下线、hello、心跳：更新运行时状态，补全 Agent 信息，并推送给 SSE。</summary>
public sealed class MonitorNodeEventSink : INodeEventSink
{
    private readonly IAgentStore _agents;
    private readonly AgentRuntime _runtime;
    private readonly IEventStore _events;
    private readonly LiveHub _live;
    private readonly AgentService _service;

    public MonitorNodeEventSink(IAgentStore agents, AgentRuntime runtime, IEventStore events, LiveHub live, AgentService service)
    {
        _agents = agents;
        _runtime = runtime;
        _events = events;
        _live = live;
        _service = service;
    }

    public async ValueTask OnEventAsync(NodeEvent nodeEvent, CancellationToken ct)
    {
        NodeId id = nodeEvent.Node.Id;
        switch (nodeEvent.Kind)
        {
            case NodeEventKind.Online:
                _live.Publish(new NodeSignal("agent", id.Value, true, nodeEvent.AtMs));
                await _events.AddAsync(id, "agent_online", nodeEvent.AtMs, null, ct);
                break;
            case NodeEventKind.Offline:
                _runtime.MarkOffline(id);
                _live.Publish(new NodeSignal("agent", id.Value, false, nodeEvent.AtMs));
                await _events.AddAsync(id, "agent_offline", nodeEvent.AtMs, null, ct);
                break;
            case NodeEventKind.Hello when nodeEvent.Hello is { } hello:
                await ApplyHelloAsync(id, hello, ct);
                break;
            case NodeEventKind.Heartbeat when nodeEvent.Extra is { } extra:
                try
                {
                    AgentRuntimeState state = _runtime.Get(id);
                    AgentHeartbeatExtra? beat = extra.Deserialize(MonitorJsonContext.Default.AgentHeartbeatExtra);
                    state.LastHeartbeat = beat;
                    state.LastHeartbeatAtMs = nodeEvent.AtMs;
                    if (beat is not null)
                    {
                        state.ReplaceProbes(beat.Probes.Select(p => new ConnectedProbeInfo(p.Id.Value, p.Version, p.Remote, p.SinceUtcMs)));
                    }
                }
                catch (JsonException)
                {
                    // 忽略格式不正确的附加字段
                }

                break;
        }
    }

    private async ValueTask ApplyHelloAsync(NodeId id, HelloParams hello, CancellationToken ct)
    {
        var kinds = new List<string>();
        foreach (string capability in hello.Capabilities)
        {
            if (capability.StartsWith(MonitorCapabilities.KindsPrefix, StringComparison.Ordinal))
            {
                kinds.AddRange(capability[MonitorCapabilities.KindsPrefix.Length..].Split(',', StringSplitOptions.RemoveEmptyEntries | StringSplitOptions.TrimEntries));
            }
        }

        AgentHelloMeta meta = new();
        if (hello.Meta is { } raw)
        {
            try
            {
                meta = raw.Deserialize(MonitorJsonContext.Default.AgentHelloMeta) ?? meta;
            }
            catch (JsonException)
            {
                // 忽略
            }
        }

        byte[]? spki = null;
        if (meta.TlsSpki is not null)
        {
            try
            {
                spki = Convert.FromBase64String(meta.TlsSpki);
            }
            catch (FormatException)
            {
                // 忽略
            }
        }

        AgentRecord? before = await _agents.GetAsync(id, ct);
        string? limitsJson = meta.LocalLimits is null ? null : JsonSerializer.Serialize(meta.LocalLimits, LocalLimitsJson.Default.LocalLimits);
        await _agents.ApplyHelloAsync(id, new AgentHello(kinds, hello.NodeVersion, limitsJson, spki is { Length: 32 } ? spki : null, meta.ProbeTcpPort, meta.ProbeUdpPorts), ct);

        // 端口或证书指纹变了，Probe 需要新的列表
        AgentRecord? after = await _agents.GetAsync(id, ct);
        if (before is not null && after is not null
            && (!before.TlsSpki.AsSpan().SequenceEqual(after.TlsSpki) || before.ProbeTcpPort != after.ProbeTcpPort || !before.ProbeUdpPorts.SequenceEqual(after.ProbeUdpPorts))
            && after.Status == NodeStatus.Active)
        {
            await _service.AfterAgentSetChangedAsync(ct);
        }
    }
}

[System.Text.Json.Serialization.JsonSerializable(typeof(LocalLimits))]
internal sealed partial class LocalLimitsJson : System.Text.Json.Serialization.JsonSerializerContext;