Fix inflated RAM usage reporting for LXC containers

Related to #553

## Problem

LXC containers showed inflated memory usage (e.g., 90%+ when actual usage was 50-60%,
96% when actual was 61%) because the code used the raw `mem` value from Proxmox's
`/cluster/resources` API endpoint. This value comes from cgroup `memory.current` which
includes reclaimable cache and buffers, making memory appear nearly full even when
plenty is available.

## Root Cause

- **Nodes**: Had sophisticated cache-aware memory calculation with RRD fallbacks
- **VMs (qemu)**: Had detailed memory calculation using guest agent meminfo
- **LXCs**: Naively used `res.Mem` directly without any cache-aware correction

The Proxmox cluster resources API's `mem` field for LXCs includes cache/buffers
(from cgroup memory accounting), which should be excluded for accurate "used" memory.

## Solution

Implement cache-aware memory calculation for LXC containers by:

1. Adding `GetLXCRRDData()` method to fetch RRD metrics for LXC containers from
   `/nodes/{node}/lxc/{vmid}/rrddata`
2. Using RRD `memavailable` to calculate actual used memory (total - available)
3. Falling back to RRD `memused` if `memavailable` is not available
4. Only using cluster resources `mem` value as last resort

This matches the approach already used for nodes and VMs, providing consistent
cache-aware memory reporting across all resource types.

## Changes

- Added `GuestRRDPoint` type and `GetLXCRRDData()` method to pkg/proxmox
- Added `GetLXCRRDData()` to ClusterClient for cluster-aware operations
- Modified LXC memory calculation in `pollPVEInstance()` to use RRD data when available
- Added guest memory snapshot recording for LXC containers
- Updated test stubs to implement the new interface method

## Testing

- Code compiles successfully
- Follows the same proven pattern used for nodes and VMs
- Includes diagnostic snapshot recording for troubleshooting

internal/monitoring/monitor.go

@@ -46,6 +46,7 @@ type PVEClientInterface interface {
 	GetNodes(ctx context.Context) ([]proxmox.Node, error)
 	GetNodeStatus(ctx context.Context, node string) (*proxmox.NodeStatus, error)
 	GetNodeRRDData(ctx context.Context, node string, timeframe string, cf string, ds []string) ([]proxmox.NodeRRDPoint, error)
+	GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe string, cf string, ds []string) ([]proxmox.GuestRRDPoint, error)
 	GetVMs(ctx context.Context, node string) ([]proxmox.VM, error)
 	GetContainers(ctx context.Context, node string) ([]proxmox.Container, error)
 	GetStorage(ctx context.Context, node string) ([]proxmox.Storage, error)
@@ -6562,15 +6563,82 @@ func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceNam
 			}
 
 			// Calculate I/O rates for container
+			sampleTime := time.Now()
 			currentMetrics := IOMetrics{
 				DiskRead:   int64(res.DiskRead),
 				DiskWrite:  int64(res.DiskWrite),
 				NetworkIn:  int64(res.NetIn),
 				NetworkOut: int64(res.NetOut),
-				Timestamp:  time.Now(),
+				Timestamp:  sampleTime,
 			}
 			diskReadRate, diskWriteRate, netInRate, netOutRate := m.rateTracker.CalculateRates(guestID, currentMetrics)
 
+			// Calculate cache-aware memory for LXC containers
+			// The cluster resources API returns mem from cgroup which includes cache/buffers (inflated).
+			// Try to get more accurate memory metrics from RRD data.
+			memTotal := res.MaxMem
+			memUsed := res.Mem
+			memorySource := "cluster-resources"
+			guestRaw := VMMemoryRaw{
+				ListingMem:    res.Mem,
+				ListingMaxMem: res.MaxMem,
+			}
+
+			// For running containers, try to get RRD data for cache-aware memory calculation
+			if res.Status == "running" {
+				rrdCtx, rrdCancel := context.WithTimeout(ctx, 5*time.Second)
+				rrdPoints, err := client.GetLXCRRDData(rrdCtx, res.Node, res.VMID, "hour", "AVERAGE", []string{"memavailable", "memused", "maxmem"})
+				rrdCancel()
+
+				if err == nil && len(rrdPoints) > 0 {
+					// Use the most recent RRD point
+					point := rrdPoints[len(rrdPoints)-1]
+
+					if point.MaxMem != nil && *point.MaxMem > 0 {
+						guestRaw.StatusMaxMem = uint64(*point.MaxMem)
+					}
+
+					// Prefer memavailable-based calculation (excludes cache/buffers)
+					if point.MemAvailable != nil && *point.MemAvailable > 0 {
+						memAvailable := uint64(*point.MemAvailable)
+						if memAvailable <= memTotal {
+							memUsed = memTotal - memAvailable
+							memorySource = "rrd-memavailable"
+							guestRaw.MemInfoAvailable = memAvailable
+							log.Debug().
+								Str("container", res.Name).
+								Str("node", res.Node).
+								Uint64("total", memTotal).
+								Uint64("available", memAvailable).
+								Uint64("used", memUsed).
+								Float64("usage", safePercentage(float64(memUsed), float64(memTotal))).
+								Msg("LXC memory: using RRD memavailable (excludes reclaimable cache)")
+						}
+					} else if point.MemUsed != nil && *point.MemUsed > 0 {
+						// Fall back to memused from RRD if available
+						memUsed = uint64(*point.MemUsed)
+						if memUsed <= memTotal {
+							memorySource = "rrd-memused"
+							guestRaw.MemInfoUsed = memUsed
+							log.Debug().
+								Str("container", res.Name).
+								Str("node", res.Node).
+								Uint64("total", memTotal).
+								Uint64("used", memUsed).
+								Float64("usage", safePercentage(float64(memUsed), float64(memTotal))).
+								Msg("LXC memory: using RRD memused (excludes reclaimable cache)")
+						}
+					}
+				} else if err != nil {
+					log.Debug().
+						Err(err).
+						Str("instance", instanceName).
+						Str("container", res.Name).
+						Int("vmid", res.VMID).
+						Msg("RRD memory data unavailable for LXC, using cluster resources value")
+				}
+			}
+
 			container := models.Container{
 				ID:       guestID,
 				VMID:     res.VMID,
@@ -6582,10 +6650,10 @@ func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceNam
 				CPU:      safeFloat(res.CPU),
 				CPUs:     int(res.MaxCPU),
 				Memory: models.Memory{
-					Total: int64(res.MaxMem),
+					Total: int64(memTotal),
-					Used:  int64(res.Mem),
+					Used:  int64(memUsed),
-					Free:  int64(res.MaxMem - res.Mem),
+					Free:  int64(memTotal - memUsed),
-					Usage: safePercentage(float64(res.Mem), float64(res.MaxMem)),
+					Usage: safePercentage(float64(memUsed), float64(memTotal)),
 				},
 				Disk: models.Disk{
 					Total: int64(res.MaxDisk),
@@ -6623,6 +6691,15 @@ func (m *Monitor) pollVMsAndContainersEfficient(ctx context.Context, instanceNam
 
 			allContainers = append(allContainers, container)
 
+			m.recordGuestSnapshot(instanceName, container.Type, res.Node, res.VMID, GuestMemorySnapshot{
+				Name:         container.Name,
+				Status:       container.Status,
+				RetrievedAt:  sampleTime,
+				MemorySource: memorySource,
+				Memory:       container.Memory,
+				Raw:          guestRaw,
+			})
+
 			// For non-running containers, zero out resource usage metrics to prevent false alerts
 			// Proxmox may report stale or residual metrics for stopped containers
 			if container.Status != "running" {

internal/monitoring/monitor_memory_test.go

@@ -33,6 +33,10 @@ func (s *stubPVEClient) GetNodeRRDData(ctx context.Context, node, timeframe, cf
 	return s.rrdPoints, nil
 }
 
+func (s *stubPVEClient) GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe, cf string, ds []string) ([]proxmox.GuestRRDPoint, error) {
+	return nil, nil
+}
+
 func (s *stubPVEClient) GetVMs(ctx context.Context, node string) ([]proxmox.VM, error) {
 	return nil, nil
 }

internal/monitoring/monitor_snapshots_test.go

@@ -20,6 +20,9 @@ func (f fakeSnapshotClient) GetNodeStatus(ctx context.Context, node string) (*pr
 func (f fakeSnapshotClient) GetNodeRRDData(ctx context.Context, node string, timeframe string, cf string, ds []string) ([]proxmox.NodeRRDPoint, error) {
 	return nil, nil
 }
+func (f fakeSnapshotClient) GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe string, cf string, ds []string) ([]proxmox.GuestRRDPoint, error) {
+	return nil, nil
+}
 func (f fakeSnapshotClient) GetVMs(ctx context.Context, node string) ([]proxmox.VM, error) {
 	return nil, nil
 }

internal/monitoring/monitor_storage_test.go

@@ -30,6 +30,10 @@ func (f *fakeStorageClient) GetNodeRRDData(ctx context.Context, node string, tim
 	return nil, nil
 }
 
+func (f *fakeStorageClient) GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe string, cf string, ds []string) ([]proxmox.GuestRRDPoint, error) {
+	return nil, nil
+}
+
 func (f *fakeStorageClient) GetVMs(ctx context.Context, node string) ([]proxmox.VM, error) {
 	return nil, nil
 }

pkg/proxmox/client.go

@@ -467,6 +467,14 @@ type NodeRRDPoint struct {
 	MemAvailable *float64 `json:"memavailable,omitempty"`
 }
 
+// GuestRRDPoint represents a single RRD datapoint for a VM or LXC container.
+type GuestRRDPoint struct {
+	Time         int64    `json:"time"`
+	MaxMem       *float64 `json:"maxmem,omitempty"`
+	MemUsed      *float64 `json:"memused,omitempty"`
+	MemAvailable *float64 `json:"memavailable,omitempty"`
+}
+
 // NodeStatus represents detailed node status from /nodes/{node}/status endpoint
 // This endpoint provides real-time metrics that update every second
 type NodeStatus struct {
@@ -709,6 +717,44 @@ func (c *Client) GetNodeRRDData(ctx context.Context, node, timeframe, cf string,
 	return result.Data, nil
 }
 
+// GetLXCRRDData retrieves RRD metrics for an LXC container.
+func (c *Client) GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe, cf string, ds []string) ([]GuestRRDPoint, error) {
+	if timeframe == "" {
+		timeframe = "hour"
+	}
+	if cf == "" {
+		cf = "AVERAGE"
+	}
+
+	params := url.Values{}
+	params.Set("timeframe", timeframe)
+	params.Set("cf", cf)
+	if len(ds) > 0 {
+		params.Set("ds", strings.Join(ds, ","))
+	}
+
+	path := fmt.Sprintf("/nodes/%s/lxc/%d/rrddata", url.PathEscape(node), vmid)
+	if query := params.Encode(); query != "" {
+		path = fmt.Sprintf("%s?%s", path, query)
+	}
+
+	resp, err := c.get(ctx, path)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	var result struct {
+		Data []GuestRRDPoint `json:"data"`
+	}
+
+	if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
+		return nil, err
+	}
+
+	return result.Data, nil
+}
+
 // VM represents a Proxmox VE virtual machine
 type VM struct {
 	VMID      int     `json:"vmid"`

pkg/proxmox/cluster_client.go

@@ -792,6 +792,19 @@ func (cc *ClusterClient) GetNodeRRDData(ctx context.Context, node, timeframe, cf
 	return result, err
 }
 
+func (cc *ClusterClient) GetLXCRRDData(ctx context.Context, node string, vmid int, timeframe, cf string, ds []string) ([]GuestRRDPoint, error) {
+	var result []GuestRRDPoint
+	err := cc.executeWithFailover(ctx, func(client *Client) error {
+		points, err := client.GetLXCRRDData(ctx, node, vmid, timeframe, cf, ds)
+		if err != nil {
+			return err
+		}
+		result = points
+		return nil
+	})
+	return result, err
+}
+
 func (cc *ClusterClient) GetVMs(ctx context.Context, node string) ([]VM, error) {
 	var result []VM
 	err := cc.executeWithFailover(ctx, func(client *Client) error {