Files
at-container-registry/pkg/hold/admin/handlers.go
T
Evan JarrettandClaude Opus 5 7d9de7c090 admin: resolve handles after limiting top users
The dashboard's top-users panel resolved a handle for every user with a
quota record, then sorted and truncated to ten. On a hold with ~500 crew
that is ~500 serial identity lookups to render ten rows.

Each lookup goes through the shared identity directory, whose HTTP client
allows 10s per request. One stalled lookup consumed the entire reverse
proxy budget, so the panel returned a partial body and the client hung up
mid-render:

  admin/auth.go:161 "Failed to render template"
    template=partials/top_users.html
    error="write: broken pipe"
  "GET /admin/api/top-users?limit=10" - 200 4096B in 10.005s

Sort and truncate first, then resolve only the surviving rows, so the
count is bounded by the limit rather than by hold size. Resolve those
concurrently under a 3s deadline: a slow lookup now degrades to a bare
DID instead of taking the whole request down with it.

The crew tab has the same underlying problem in a different shape, one
lazy-load request per row, and is not addressed here.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-09 20:22:46 -05:00

213 lines
5.6 KiB
Go

package admin
import (
"context"
"log/slog"
"net/http"
"sort"
"strconv"
"sync"
"time"
"atcr.io/pkg/atproto"
"github.com/go-chi/render"
)
// topUsersResolveTimeout bounds handle resolution on the dashboard's top-users
// panel. The shared identity directory allows each lookup 10s, which is the
// whole reverse proxy budget, so the panel caps its own resolution well short
// of that and renders bare DIDs for anything slower.
const topUsersResolveTimeout = 3 * time.Second
// DashboardStats contains dashboard statistics
type DashboardStats struct {
TotalCrewMembers int
TierDistribution map[string]int
}
// handleAdmin renders the single admin page with client-side tab switching
func (ui *AdminUI) handleAdmin(w http.ResponseWriter, r *http.Request) {
defer clearFlash(w)
ui.renderTemplate(w, "pages/admin.html", ui.newPageData(r, "Hold Admin", ""))
}
// getDashboardStats returns dashboard statistics
func (ui *AdminUI) getDashboardStats(ctx context.Context) DashboardStats {
crew, err := ui.pds.ListCrewMembers(ctx)
if err != nil {
slog.Warn("Failed to list crew members for dashboard", "error", err)
}
stats := DashboardStats{
TotalCrewMembers: len(crew),
TierDistribution: make(map[string]int),
}
defaultTier := "default"
if ui.quotaMgr != nil && ui.quotaMgr.IsEnabled() {
defaultTier = ui.quotaMgr.GetDefaultTier()
}
for _, member := range crew {
tier := member.Record.Tier
if tier == "" {
tier = defaultTier
}
stats.TierDistribution[tier]++
}
return stats
}
// handleDashboardTab returns the dashboard tab content (HTMX partial)
func (ui *AdminUI) handleDashboardTab(w http.ResponseWriter, r *http.Request) {
defer clearFlash(w)
data := struct {
Stats DashboardStats
}{
Stats: ui.getDashboardStats(r.Context()),
}
ui.renderTemplate(w, "partials/tab_dashboard.html", data)
}
// StorageStats contains storage statistics
type StorageStats struct {
TotalBlobs int `json:"totalBlobs"`
TotalSize int64 `json:"totalSize"`
TotalHuman string `json:"totalHuman"`
UniqueDigests int `json:"uniqueDigests"`
}
// handleStatsAPI returns storage statistics (for HTMX lazy loading)
func (ui *AdminUI) handleStatsAPI(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
// Get layer record count from the index
recordsIndex := ui.pds.RecordsIndex()
if recordsIndex == nil {
http.Error(w, "Records index not available", http.StatusInternalServerError)
return
}
// Count total layer records
totalBlobs, err := recordsIndex.Count(atproto.LayerCollection)
if err != nil {
slog.Error("Failed to count layer records", "error", err)
http.Error(w, "Failed to load stats", http.StatusInternalServerError)
return
}
// Calculate total storage with a single bulk query
var totalSize int64
uniqueDigests := 0
allQuotas, err := ui.pds.GetAllUserQuotas(ctx)
if err != nil {
slog.Warn("Failed to get all user quotas", "error", err)
} else {
for _, q := range allQuotas {
totalSize += q.TotalSize
uniqueDigests += q.UniqueBlobs
}
}
stats := StorageStats{
TotalBlobs: totalBlobs,
TotalSize: totalSize,
TotalHuman: formatHumanBytes(totalSize),
UniqueDigests: uniqueDigests,
}
// If HTMX request, return HTML partial
if r.Header.Get("HX-Request") == "true" {
data := struct {
Stats StorageStats
}{Stats: stats}
ui.renderTemplate(w, "partials/usage_stats.html", data)
return
}
// Otherwise return JSON
render.JSON(w, r, stats)
}
// UserUsage represents storage usage for a user
type UserUsage struct {
DID string `json:"did"`
Handle string `json:"handle"`
Usage int64 `json:"usage"`
UsageHuman string `json:"usageHuman"`
BlobCount int `json:"blobCount"`
}
// handleTopUsersAPI returns top users by storage (for HTMX lazy loading)
func (ui *AdminUI) handleTopUsersAPI(w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
limit := 10
if l := r.URL.Query().Get("limit"); l != "" {
if parsed, err := strconv.Atoi(l); err == nil && parsed > 0 && parsed <= 100 {
limit = parsed
}
}
// Get all user quotas in a single bulk query
allQuotas, err := ui.pds.GetAllUserQuotas(ctx)
if err != nil {
slog.Error("Failed to get all user quotas", "error", err)
http.Error(w, "Failed to load top users", http.StatusInternalServerError)
return
}
var users []UserUsage
for did, q := range allQuotas {
users = append(users, UserUsage{
DID: did,
Usage: q.TotalSize,
UsageHuman: formatHumanBytes(q.TotalSize),
BlobCount: q.UniqueBlobs,
})
}
// Sort by usage (highest first)
sort.Slice(users, func(i, j int) bool {
return users[i].Usage > users[j].Usage
})
// Limit results
if len(users) > limit {
users = users[:limit]
}
// Resolve handles only for the rows that survived the limit. Doing this
// inside the loop above cost one network lookup per user on the hold to
// display ten of them, which blew past the identity directory's 10s HTTP
// timeout and left the client hanging up mid-render. Resolve concurrently
// under a deadline well inside the reverse proxy's budget so a stalled
// lookup degrades to a bare DID instead of a dead request.
resolveCtx, cancel := context.WithTimeout(ctx, topUsersResolveTimeout)
defer cancel()
var wg sync.WaitGroup
for i := range users {
wg.Add(1)
go func(u *UserUsage) {
defer wg.Done()
u.Handle = resolveHandle(resolveCtx, u.DID)
}(&users[i])
}
wg.Wait()
// If HTMX request, return HTML partial
if r.Header.Get("HX-Request") == "true" {
data := struct {
Users []UserUsage
}{Users: users}
ui.renderTemplate(w, "partials/top_users.html", data)
return
}
// Otherwise return JSON
render.JSON(w, r, users)
}