feat(runtime): track node lifecycle and QUIC link state

backend/src/MoShiro.Backend.Api/Admin/LinkEndpoints.cs

@@ -20,9 +20,12 @@ public static class LinkEndpoints
         group.MapPut("/links/{probe_id}/{agent_id}", SetAsync)
             .Api("links_set", "Force a link on or off, or give it back to the automatic setting (enabled: null)", Tag).Returns<LinkDto>().ProducesProblems(400, 401, 403, 404)
             .Requires(Capability.FleetWrite);
+        group.MapPost("/links/{probe_id}/{agent_id}/retry", RetryAsync)
+            .Api("links_retry", "Enable and immediately retry the Probe-to-Agent connection", Tag).Returns<LinkRetryDto>().ProducesProblems(401, 403, 404)
+            .Requires(Capability.FleetWrite);
     }
 
-    private static async Task<IResult> ListAsync(AgentService agents, ProbeService probes, LinkService links, CancellationToken ct)
+    private static async Task<IResult> ListAsync(AgentService agents, ProbeService probes, LinkService links, ProbeQuicRuntime quic, TimeProvider time, CancellationToken ct)
     {
         LinkSnapshot snapshot = await links.SnapshotAsync(ct);
         IReadOnlyList<AgentView> agentViews = await agents.ListAsync(ct);
@@ -31,14 +34,14 @@ public static class LinkEndpoints
         {
             foreach (AgentView agent in agentViews.OrderBy(a => a.Agent.Name, StringComparer.Ordinal))
             {
-                items.Add(ToDto(snapshot, probe, agent));
+                items.Add(ToDto(snapshot, probe, agent, quic, time));
             }
         }
 
         return TypedResults.Json(new LinkMatrixDto(snapshot.Auto, items), ApiJsonContext.Default.LinkMatrixDto);
     }
 
-    private static async Task<IResult> SetAsync(HttpContext http, AgentService agents, ProbeService probes, LinkService links, string probe_id, string agent_id, SetLinkRequest request, CancellationToken ct)
+    private static async Task<IResult> SetAsync(HttpContext http, AgentService agents, ProbeService probes, LinkService links, ProbeQuicRuntime quic, TimeProvider time, string probe_id, string agent_id, SetLinkRequest request, CancellationToken ct)
     {
         if (await probes.GetAsync(probe_id, ct) is not { Revoked: false } probe)
         {
@@ -52,10 +55,71 @@ public static class LinkEndpoints
 
         await links.SetAsync(probe_id, agent_id, request.Enabled, ct);
         await agents.AfterAgentSetChangedAsync(ct); // 重新下发：Probe 的 Agent 列表、Agent 允许的 Probe
-        return TypedResults.Json(ToDto(await links.SnapshotAsync(ct), probe, agent), ApiJsonContext.Default.LinkDto);
+        if (request.Enabled == true)
+        {
+            await agents.NotifyProbeLinkRetryAsync(new NodeId(probe_id), agent.Agent.Id, ct);
+        }
+        return TypedResults.Json(ToDto(await links.SnapshotAsync(ct), probe, agent, quic, time), ApiJsonContext.Default.LinkDto);
     }
 
-    private static LinkDto ToDto(LinkSnapshot snapshot, ProbeView probe, AgentView agent) =>
-        new(probe.Probe.Id, probe.Probe.Name, agent.Agent.Id.Value, agent.Agent.Name, snapshot.IsLinked(probe.Probe.Id, agent.Agent.Id.Value), snapshot.OverrideOf(probe.Probe.Id, agent.Agent.Id.Value),
-            agent.Probes.Any(p => p.Id == probe.Probe.Id));
+    private static async Task<IResult> RetryAsync(HttpContext http, AgentService agents, ProbeService probes, LinkService links, ProbeQuicRuntime quic, TimeProvider time, string probe_id, string agent_id, CancellationToken ct)
+    {
+        if (await probes.GetAsync(probe_id, ct) is not { Revoked: false } probe)
+        {
+            return Problems.Create(http, MoShiroErrors.ProbeNotFound, "probe not found");
+        }
+        if (await agents.GetAsync(new NodeId(agent_id), ct) is not { } agent)
+        {
+            return Problems.Create(http, MoShiroErrors.AgentNotFound, "agent not found");
+        }
+
+        await links.SetAsync(probe_id, agent_id, true, ct);
+        await agents.AfterAgentSetChangedAsync(ct);
+        bool queued = await agents.NotifyProbeLinkRetryAsync(new NodeId(probe_id), agent.Agent.Id, ct);
+        LinkDto link = ToDto(await links.SnapshotAsync(ct), probe, agent, quic, time);
+        return TypedResults.Json(new LinkRetryDto(queued, link), ApiJsonContext.Default.LinkRetryDto, statusCode: StatusCodes.Status202Accepted);
+    }
+
+    private static LinkDto ToDto(LinkSnapshot snapshot, ProbeView probe, AgentView agent, ProbeQuicRuntime quicRuntime, TimeProvider time)
+    {
+        bool linked = snapshot.IsLinked(probe.Probe.Id, agent.Agent.Id.Value);
+        ProbeQuicLinkSnapshot? quicSnapshot = quicRuntime.Get(probe.Probe.Id, agent.Agent.Id.Value, IpVersions.V4);
+        ProbeQuicLinkHealth? health = quicSnapshot?.Health;
+        bool fresh = quicSnapshot is not null && time.GetUtcNow().ToUnixTimeMilliseconds() - quicSnapshot.AtMs <= 30_000;
+        string state;
+        string? reason = null;
+        if (!linked) { state = "disabled"; reason = "线路策略已关闭"; }
+        else if (probe.RuntimeStatus != NodeRuntimeStatus.Running) { state = "probe_unavailable"; reason = $"Probe 状态：{probe.RuntimeStatus}"; }
+        else if (agent.RuntimeStatus != NodeRuntimeStatus.Running) { state = "agent_unavailable"; reason = $"Agent 状态：{agent.RuntimeStatus}"; }
+        else if (health is null || !fresh)
+        {
+            state = "waiting_report";
+            reason = health is null ? "Probe 尚未上报此线路的 QUIC 拨号状态" : "Probe 的线路状态上报已过期";
+        }
+        else if (health.Connected && agent.Probes.Any(p => p.Id == probe.Probe.Id))
+        {
+            state = "connected";
+        }
+        else if (health.State == ProbeLinkAttemptState.RetryWait)
+        {
+            state = "retry_wait";
+            reason = health.LastError is { Length: > 0 } error ? $"QUIC 连接失败：{error}" : "QUIC 连接失败，等待自动重试";
+        }
+        else if (agent.Probes.All(p => p.Id != probe.Probe.Id))
+        {
+            state = "control_disconnected";
+            reason = health.LastError ?? "Probe 尚未建立到 Agent 的测量控制会话";
+        }
+        else
+        {
+            state = "quic_connecting";
+            reason = health.LastError ?? "控制会话已建立，正在建立 QUIC 测量会话";
+        }
+        bool connected = state == "connected";
+        return new LinkDto(probe.Probe.Id, probe.Probe.Name, agent.Agent.Id.Value, agent.Agent.Name, linked,
+            snapshot.OverrideOf(probe.Probe.Id, agent.Agent.Id.Value), connected, probe.RuntimeStatus, agent.RuntimeStatus,
+            state, reason, health?.LastError, health?.RetryAttempt ?? 0, health?.RetryInMs);
+    }
 }
+
+public sealed record LinkRetryDto(bool Queued, LinkDto Link);

backend/src/MoShiro.Backend.Api/Common/Mappers.cs

@@ -46,7 +46,12 @@ public static class Mappers
             health.Connected, health.HandshakeAttempts, health.HandshakeSuccesses, health.HandshakeFailures, health.HandshakeTimeouts,
             Rate(health.HandshakeSuccesses), Rate(health.HandshakeFailures), Rate(health.HandshakeTimeouts), health.HandshakeP95Ms, health.HandshakeSampleCount,
             health.TransportPacketLossRate, health.TransportPacketsSent, health.TransportCongestionEvents,
-            health.TransportEcnCongestionEvents, health.PathMtu, Time(snapshot.AtMs));
+            health.TransportEcnCongestionEvents, health.PathMtu, Time(snapshot.AtMs), health.State switch
+            {
+                ProbeLinkAttemptState.Connected => "connected",
+                ProbeLinkAttemptState.RetryWait => "retry_wait",
+                _ => "connecting",
+            }, health.RetryAttempt, health.RetryInMs);
     }
 
     public static HostDto ToDto(HostSnapshot h) =>
@@ -57,7 +62,7 @@ public static class Mappers
     public static ProbeDto ToDto(ProbeView v) =>
         new(
             v.Probe.Id, v.Probe.Name, Time(v.Probe.CreatedAtMs), Time(v.Probe.RevokedAtMs), v.Revoked, v.Online,
-            [.. v.Connections.Select(c => new ProbeConnectionDto(c.AgentId.Value, c.AgentName, c.AgentOnline, c.Version, c.Remote, Time(c.SinceMs), c.IpVersion))], ProbeIpv6Of(v.Probe), v.Probe.AddedBy);
+            [.. v.Connections.Select(c => new ProbeConnectionDto(c.AgentId.Value, c.AgentName, c.AgentOnline, c.Version, c.Remote, Time(c.SinceMs), c.IpVersion))], v.RuntimeStatus, ProbeIpv6Of(v.Probe), v.Probe.AddedBy);
 
     public static SeriesDto ToDto(SeriesResult r, bool live) =>
         new(
@@ -119,7 +124,7 @@ public static class Mappers
             : null;
         return new AgentDto(
             a.Id.Value, a.Name, StatusName(a.Status), view.Online, view.ProbeConnected, [.. view.Probes.Select(ToDto)], view.Host is null ? null : ToDto(view.Host), a.PublicHost, a.ProbeTcpPort, a.ProbeUdpPorts, a.ProbeUdpPorts, a.Labels, a.PublicVisible,
-            a.Kinds, a.AgentVersion, limits, a.TaskVersion, a.AppliedTaskVersion, rejected, heartbeat, view.EndpointChanges, Time(a.CreatedAtMs), Time(a.LastSeenAtMs), a.Ipv6Host, a.Ipv6Supported, Ipv6Of(view), !a.PublicHostManual, !a.Ipv6HostManual, a.AddedBy);
+            a.Kinds, a.AgentVersion, limits, a.TaskVersion, a.AppliedTaskVersion, rejected, heartbeat, view.EndpointChanges, view.RuntimeStatus, Time(a.CreatedAtMs), Time(a.LastSeenAtMs), a.Ipv6Host, a.Ipv6Supported, Ipv6Of(view), !a.PublicHostManual, !a.Ipv6HostManual, a.AddedBy);
     }
 
     public static Ipv6DiagnosticDto Ipv6Of(AgentView view)

backend/src/MoShiro.Backend.Api/Contracts/AdminDtos.cs

@@ -1,5 +1,6 @@
 using Fleet.Core;
 using System.Text.Json;
+using MoShiro.Protocol;
 
 namespace MoShiro.Backend.Api;
 
@@ -53,6 +54,7 @@ public sealed record AgentDto(
     IReadOnlyList<RejectedDto> Rejected,
     HeartbeatDto? Heartbeat,
     int EndpointChanges,
+    NodeRuntimeStatus RuntimeStatus,
     DateTimeOffset CreatedAt,
     DateTimeOffset? LastSeenAt,
     string Ipv6Host = "",
@@ -123,10 +125,10 @@ public sealed record ConnectedProbeDto(string ProbeId, string? Version, string?
 /// <summary>某个 Probe 现在连着的一个 Agent。</summary>
 public sealed record ProbeConnectionDto(string AgentId, string AgentName, bool AgentOnline, string? Version, string? Remote, DateTimeOffset ConnectedSince, string IpVersion = "ipv4");
 
-public sealed record ProbeDto(string Id, string Name, DateTimeOffset CreatedAt, DateTimeOffset? RevokedAt, bool Revoked, bool Online, IReadOnlyList<ProbeConnectionDto> Connections, Ipv6DiagnosticDto? Ipv6 = null, string? AddedBy = null);
+public sealed record ProbeDto(string Id, string Name, DateTimeOffset CreatedAt, DateTimeOffset? RevokedAt, bool Revoked, bool Online, IReadOnlyList<ProbeConnectionDto> Connections, NodeRuntimeStatus RuntimeStatus, Ipv6DiagnosticDto? Ipv6 = null, string? AddedBy = null);
 
 /// <summary>一对（Probe，Agent）的线路开关：<c>linked</c> 是最终结果，<c>override</c> 是手动设置（<c>null</c> = 跟随自动连接），<c>connected</c> 是现在有没有连着。</summary>
-public sealed record LinkDto(string ProbeId, string ProbeName, string AgentId, string AgentName, bool Linked, bool? Override, bool Connected);
+public sealed record LinkDto(string ProbeId, string ProbeName, string AgentId, string AgentName, bool Linked, bool? Override, bool Connected, NodeRuntimeStatus ProbeRuntimeStatus, NodeRuntimeStatus AgentRuntimeStatus, string ConnectionState = "connecting", string? ConnectionReason = null, string? LastError = null, int RetryAttempt = 0, long? RetryInMs = null);
 
 public sealed record LinkMatrixDto(bool AutoLink, IReadOnlyList<LinkDto> Items);
 

backend/src/MoShiro.Backend.Api/Contracts/ApiJsonContext.cs

@@ -41,6 +41,7 @@ namespace MoShiro.Backend.Api;
 [JsonSerializable(typeof(AgentPatchRequest))]
 [JsonSerializable(typeof(LinkDto))]
 [JsonSerializable(typeof(LinkMatrixDto))]
+[JsonSerializable(typeof(LinkRetryDto))]
 [JsonSerializable(typeof(SetLinkRequest))]
 [JsonSerializable(typeof(BuildDto))]
 [JsonSerializable(typeof(ReleaseDto))]

backend/src/MoShiro.Backend.Api/Contracts/PublicDtos.cs

@@ -48,7 +48,10 @@ public sealed record QuicLineHealthDto(
     uint? TransportCongestionEvents,
     uint? TransportEcnCongestionEvents,
     ushort? PathMtu,
-    DateTimeOffset At);
+    DateTimeOffset At,
+    string State = "connecting",
+    int RetryAttempt = 0,
+    long? RetryInMs = null);
 
 public sealed record OverviewDto(string Range, DateTimeOffset From, DateTimeOffset To, string Step, string Source, bool Degraded, IReadOnlyList<OverviewLineDto> Lines);
 

backend/src/MoShiro.Backend.Application/Agents/AgentRuntime.cs

@@ -13,6 +13,7 @@ public sealed class AgentRuntimeState
     private readonly object _gate = new();
     private int _endpointChanges;
     private IReadOnlyList<ConnectedProbeInfo> _probes = [];
+    private string? _lastDisconnectReason;
 
     /// <summary>当前连着这个 Agent 的 Probe（每个 Probe 一项）。</summary>
     public IReadOnlyList<ConnectedProbeInfo> Probes => Volatile.Read(ref _probes);
@@ -60,6 +61,17 @@ public sealed class AgentRuntimeState
 
     public int EndpointChanges => Volatile.Read(ref _endpointChanges);
 
+    public string? LastDisconnectReason => Volatile.Read(ref _lastDisconnectReason);
+
+    public void MarkOnline()
+    {
+        Volatile.Write(ref _lastDisconnectReason, null);
+        LastHeartbeat = null;
+        LastHeartbeatAtMs = 0;
+    }
+
+    public void MarkOffline(string? reason) => Volatile.Write(ref _lastDisconnectReason, reason);
+
     public void CountEndpointChange() => Interlocked.Increment(ref _endpointChanges);
 }
 

backend/src/MoShiro.Backend.Application/Agents/AgentService.cs

@@ -7,10 +7,46 @@ using MoShiro.Protocol;
 
 namespace MoShiro.Backend.Application;
 
-public sealed record AgentView(AgentRecord Agent, bool Online, IReadOnlyList<ConnectedProbeInfo> Probes, AgentHeartbeatExtra? Heartbeat, int EndpointChanges, long? HeartbeatAtMs = null, HostSnapshot? Host = null)
+public sealed record AgentView(AgentRecord Agent, bool Online, IReadOnlyList<ConnectedProbeInfo> Probes, AgentHeartbeatExtra? Heartbeat, int EndpointChanges, long? HeartbeatAtMs = null, HostSnapshot? Host = null, string? LastDisconnectReason = null)
 {
     /// <summary>至少有一个 Probe 连着。</summary>
     public bool ProbeConnected => Probes.Count > 0;
+
+    public NodeRuntimeStatus RuntimeStatus
+    {
+        get
+        {
+            if (Agent.Status == NodeStatus.Disabled)
+            {
+                return NodeRuntimeStatus.Maintenance;
+            }
+            if (Agent.Status == NodeStatus.Pending)
+            {
+                return NodeRuntimeStatus.Initializing;
+            }
+            if (!Online)
+            {
+                if (LastDisconnectReason == "updating")
+                {
+                    return NodeRuntimeStatus.Updating;
+                }
+                if (LastDisconnectReason == "stopping")
+                {
+                    return Heartbeat?.Update?.State == "applying" ? NodeRuntimeStatus.Updating : NodeRuntimeStatus.Stopping;
+                }
+                return NodeRuntimeStatus.Offline;
+            }
+            if (Heartbeat?.Update?.State == "applying")
+            {
+                return NodeRuntimeStatus.Updating;
+            }
+            if (Heartbeat is null)
+            {
+                return NodeRuntimeStatus.Initializing;
+            }
+            return Heartbeat.RuntimeStatus;
+        }
+    }
 }
 
 /// <summary>Agent 的管理：列表、修改、审批、启停、接管、删除，以及 Probe 需要的 Agent 列表。</summary>
@@ -53,9 +89,12 @@ public sealed partial class AgentService
     public AgentView View(AgentRecord agent)
     {
         _runtime.TryGet(agent.Id, out AgentRuntimeState? state);
-        return new AgentView(agent, _gateway.IsOnline(agent.Id), state?.Probes ?? [], state?.LastHeartbeat, state?.EndpointChanges ?? 0, state?.LastHeartbeatAtMs, state?.Host);
+        return new AgentView(agent, _gateway.IsOnline(agent.Id), state?.Probes ?? [], state?.LastHeartbeat, state?.EndpointChanges ?? 0, state?.LastHeartbeatAtMs, state?.Host, state?.LastDisconnectReason);
     }
 
+    public ValueTask<bool> NotifyProbeLinkRetryAsync(NodeId probe, NodeId agent, CancellationToken ct) =>
+        _gateway.NotifyProbeLinkRetryAsync(probe, agent, ct);
+
     private static bool ValidHost(string host, string ipVersion)
     {
         if (IPAddress.TryParse(host, out IPAddress? address))

backend/src/MoShiro.Backend.Application/Agents/MoShiroFleetHooks.cs

@@ -71,11 +71,13 @@ public sealed class MoShiroNodeEventSink : INodeEventSink
     private readonly IProbeStore _probes;
     private readonly NodeOwnerService _owners;
     private readonly ProbeQuicRuntime _quicRuntime;
+    private readonly ProbeRuntime _probeRuntime;
 
-    public MoShiroNodeEventSink(IAgentStore agents, AgentRuntime runtime, IEventStore events, LiveHub live, AgentService service, NodeInfoService nodes, ReleaseService releases, ScheduleService schedule, IProbeStore probes, NodeOwnerService owners, ProbeQuicRuntime quicRuntime)
+    public MoShiroNodeEventSink(IAgentStore agents, AgentRuntime runtime, IEventStore events, LiveHub live, AgentService service, NodeInfoService nodes, ReleaseService releases, ScheduleService schedule, IProbeStore probes, NodeOwnerService owners, ProbeQuicRuntime quicRuntime, ProbeRuntime probeRuntime)
     {
         _owners = owners;
         _quicRuntime = quicRuntime;
+        _probeRuntime = probeRuntime;
         _releases = releases;
         _schedule = schedule;
         _probes = probes;
@@ -99,10 +101,12 @@ public sealed class MoShiroNodeEventSink : INodeEventSink
         switch (nodeEvent.Kind)
         {
             case NodeEventKind.Online:
+                _runtime.Get(id).MarkOnline();
                 _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.Get(id).MarkOffline(nodeEvent.Reason);
                 _runtime.MarkOffline(id);
                 _releases.RecordStatus(id, null);
                 _live.Publish(new NodeSignal("agent", id.Value, false, nodeEvent.AtMs));
@@ -143,6 +147,7 @@ public sealed class MoShiroNodeEventSink : INodeEventSink
         switch (nodeEvent.Kind)
         {
             case NodeEventKind.Online:
+                _probeRuntime.MarkOnline(id.Value);
                 _live.Publish(new NodeSignal("probe", id.Value, true, nodeEvent.AtMs));
                 break;
             case NodeEventKind.Hello:
@@ -170,6 +175,7 @@ public sealed class MoShiroNodeEventSink : INodeEventSink
             case NodeEventKind.Offline:
                 _live.Publish(new NodeSignal("probe", id.Value, false, nodeEvent.AtMs));
                 _releases.RecordStatus(id, null);
+                _probeRuntime.MarkOffline(id.Value, nodeEvent.Reason);
                 _quicRuntime.MarkProbeOffline(id.Value, nodeEvent.AtMs);
                 break;
             case NodeEventKind.Heartbeat when nodeEvent.Extra is { } extra:
@@ -177,6 +183,10 @@ public sealed class MoShiroNodeEventSink : INodeEventSink
                 {
                     ProbeHeartbeatExtra? heartbeat = extra.Deserialize(MoShiroJsonContext.Default.ProbeHeartbeatExtra);
                     _releases.RecordStatus(id, heartbeat?.Update);
+                    if (heartbeat is not null)
+                    {
+                        _probeRuntime.RecordHeartbeat(id.Value, nodeEvent.AtMs, heartbeat);
+                    }
                     if (heartbeat?.QuicLinks is { } quicLinks)
                     {
                         _quicRuntime.Replace(id.Value, nodeEvent.AtMs, quicLinks);

backend/src/MoShiro.Backend.Application/Agents/ProbeQuicRuntime.cs

@@ -39,6 +39,10 @@ public sealed class ProbeQuicRuntime
                 AgentId = old.AgentId,
                 IpVersion = old.IpVersion,
                 Connected = false,
+                State = ProbeLinkAttemptState.RetryWait,
+                LastError = "Probe 与 Backend 的连接中断",
+                RetryAttempt = old.RetryAttempt,
+                RetryInMs = old.RetryInMs,
                 HandshakeAttempts = old.HandshakeAttempts,
                 HandshakeSuccesses = old.HandshakeSuccesses,
                 HandshakeFailures = old.HandshakeFailures,

backend/src/MoShiro.Backend.Application/Agents/ProbeRuntime.cs

@@ -0,0 +1,22 @@
+using System.Collections.Concurrent;
+using MoShiro.Protocol;
+
+namespace MoShiro.Backend.Application;
+
+/// <summary>Volatile runtime status reported by each Probe's heartbeat and Fleet connection lifecycle.</summary>
+public sealed class ProbeRuntime
+{
+    private readonly ConcurrentDictionary<string, ProbeRuntimeSnapshot> _probes = new(StringComparer.Ordinal);
+
+    public ProbeRuntimeSnapshot? Get(string probeId) => _probes.TryGetValue(probeId, out ProbeRuntimeSnapshot? snapshot) ? snapshot : null;
+
+    public void RecordHeartbeat(string probeId, long atMs, ProbeHeartbeatExtra heartbeat) =>
+        _probes[probeId] = new ProbeRuntimeSnapshot(heartbeat, atMs, null);
+
+    public void MarkOnline(string probeId) => _probes[probeId] = new ProbeRuntimeSnapshot(null, 0, null);
+
+    public void MarkOffline(string probeId, string? reason) =>
+        _probes.AddOrUpdate(probeId, _ => new ProbeRuntimeSnapshot(null, 0, reason), (_, current) => current with { DisconnectReason = reason });
+}
+
+public sealed record ProbeRuntimeSnapshot(ProbeHeartbeatExtra? Heartbeat, long AtMs, string? DisconnectReason);

backend/src/MoShiro.Backend.Application/Agents/ProbeService.cs

@@ -8,9 +8,41 @@ namespace MoShiro.Backend.Application;
 /// <summary>某个 Probe 现在连着的一个 Agent（Agent 上报）。</summary>
 public sealed record ProbeConnection(NodeId AgentId, string AgentName, bool AgentOnline, string? Version, string? Remote, long SinceMs, string IpVersion = IpVersions.V4);
 
-public sealed record ProbeView(ProbeRecord Probe, bool Online, IReadOnlyList<ProbeConnection> Connections)
+public sealed record ProbeView(ProbeRecord Probe, bool Online, IReadOnlyList<ProbeConnection> Connections, ProbeRuntimeSnapshot? Runtime = null)
 {
     public bool Revoked => Probe.RevokedAtMs is not null;
+
+    public NodeRuntimeStatus RuntimeStatus
+    {
+        get
+        {
+            if (Revoked)
+            {
+                return NodeRuntimeStatus.Maintenance;
+            }
+            if (!Online)
+            {
+                if (Runtime?.DisconnectReason == "updating")
+                {
+                    return NodeRuntimeStatus.Updating;
+                }
+                if (Runtime?.DisconnectReason == "stopping")
+                {
+                    return Runtime.Heartbeat?.Update?.State == "applying" ? NodeRuntimeStatus.Updating : NodeRuntimeStatus.Stopping;
+                }
+                return NodeRuntimeStatus.Offline;
+            }
+            if (Runtime?.Heartbeat?.Update?.State == "applying")
+            {
+                return NodeRuntimeStatus.Updating;
+            }
+            if (Runtime?.Heartbeat is null)
+            {
+                return NodeRuntimeStatus.Initializing;
+            }
+            return Runtime.Heartbeat.RuntimeStatus;
+        }
+    }
 }
 
 /// <summary>
@@ -28,13 +60,15 @@ public sealed partial class ProbeService
     private readonly TimeProvider _time;
     private readonly NodeInfoService _nodes;
     private readonly LinkService _links;
+    private readonly ProbeRuntime _probeRuntime;
 
     private readonly NodeOwnerService _owners;
 
-    public ProbeService(NodeOwnerService owners, IProbeStore probes, IAgentStore agents, AgentRuntime runtime, IAgentGateway gateway, TimeProvider time, NodeInfoService nodes, LinkService links)
+    public ProbeService(NodeOwnerService owners, IProbeStore probes, IAgentStore agents, AgentRuntime runtime, IAgentGateway gateway, TimeProvider time, NodeInfoService nodes, LinkService links, ProbeRuntime probeRuntime)
     {
         _owners = owners;
         _links = links;
+        _probeRuntime = probeRuntime;
         _nodes = nodes;
         _probes = probes;
         _agents = agents;
@@ -133,7 +167,7 @@ public sealed partial class ProbeService
             }
         }
 
-        return new ProbeView(probe, _gateway.IsOnline(new NodeId(probe.Id)), [.. connections.OrderBy(c => c.AgentName, StringComparer.Ordinal)]);
+        return new ProbeView(probe, _gateway.IsOnline(new NodeId(probe.Id)), [.. connections.OrderBy(c => c.AgentName, StringComparer.Ordinal)], _probeRuntime.Get(probe.Id));
     }
 
     [GeneratedRegex(@"^[\p{L}\p{M}\p{N}._-]{1,64}$")]

backend/src/MoShiro.Backend.Application/ApplicationServiceCollectionExtensions.cs

@@ -16,6 +16,7 @@ public static class ApplicationServiceCollectionExtensions
         services.AddHostedService<StartupTasks>();
 
         services.AddSingleton<AgentRuntime>();
+        services.AddSingleton<ProbeRuntime>();
         services.AddSingleton<ProbeQuicRuntime>();
         services.AddSingleton<LiveHub>();
         services.AddSingleton<HotLayer>();

backend/src/MoShiro.Backend.Application/Ports/IAgentGateway.cs

@@ -30,6 +30,9 @@ public interface IAgentGateway
 
     ValueTask DisconnectAsync(NodeId agent, string reason, CancellationToken ct);
 
+    /// <summary>Notify an online Probe to immediately retry all IP paths to an Agent.</summary>
+    ValueTask<bool> NotifyProbeLinkRetryAsync(NodeId probe, NodeId agent, CancellationToken ct) => ValueTask.FromResult(false);
+
     /// <summary>把发布清单和签名原样交给节点（通知，不等回复）。</summary>
     ValueTask NotifyUpdateOfferAsync(NodeId node, string manifest, string signature, CancellationToken ct);
 

backend/src/MoShiro.Backend.Gateway/FleetGateway.cs

@@ -64,6 +64,9 @@ public sealed class FleetGateway : IAgentGateway
 
     public ValueTask DisconnectAsync(NodeId agent, string reason, CancellationToken ct) => Connections.DisconnectAsync(agent, reason, ct);
 
+    public ValueTask<bool> NotifyProbeLinkRetryAsync(NodeId probe, NodeId agent, CancellationToken ct) =>
+        Connections.NotifyAsync(probe, MoShiroMethods.ProbeLinkRetry, new ProbeLinkRetryParams { AgentId = agent }, MoShiroJsonContext.Default.ProbeLinkRetryParams, ct);
+
     public async ValueTask NotifyUpdateOfferAsync(NodeId node, string manifest, string signature, CancellationToken ct) =>
         await Connections.NotifyAsync(node, MoShiroMethods.UpdateOffer, new UpdateOfferParams { Manifest = manifest, Signature = signature }, MoShiroJsonContext.Default.UpdateOfferParams, ct);