| 1 | package metrics |
| 2 | |
| 3 | import ( |
| 4 | "context" |
| 5 | "encoding/json" |
| 6 | "errors" |
| 7 | "os" |
| 8 | "strings" |
| 9 | "time" |
| 10 | |
| 11 | "github.com/carlosframework/platform/internal/release" |
| 12 | ) |
| 13 | |
| 14 | // hourFormat is the UTC hour format used both in RawChunk.Hour and in raw |
| 15 | // chunk keys: 2006-01-02-15. |
| 16 | const hourFormat = "2006-01-02-15" |
| 17 | |
| 18 | // HourOf formats t as the UTC hour bucket a raw chunk belongs to. |
| 19 | func HourOf(t time.Time) string { |
| 20 | return t.UTC().Format(hourFormat) |
| 21 | } |
| 22 | |
| 23 | // RawKey is the bucket key for a box's raw chunk covering the given hour |
| 24 | // (itself formatted like HourOf's output, e.g. 2026-08-02-21). |
| 25 | func RawKey(box, hour string) string { |
| 26 | return "metrics/boxes/" + box + "/raw/" + hour + ".json" |
| 27 | } |
| 28 | |
| 29 | // DayKey is the bucket key for a box's rolled-up day chunk (date formatted |
| 30 | // 2006-01-02). |
| 31 | func DayKey(box, date string) string { |
| 32 | return "metrics/boxes/" + box + "/days/" + date + ".json" |
| 33 | } |
| 34 | |
| 35 | // WriteRaw marshals c indented and writes it to its hour-derived key, |
| 36 | // last-writer-wins. A box appends samples to the same hour's chunk |
| 37 | // repeatedly as the hour progresses, so a bare create-only Put (ifMatch "") |
| 38 | // would fail every write after the first with ErrCASConflict — this mirrors |
| 39 | // status.WriteBox's read-etag-then-Put dance to express "replace whatever is |
| 40 | // there, or create it" on top of release.Store's CAS-only Put: read the |
| 41 | // current etag (a not-found just means "absent", so ifMatch stays "" for a |
| 42 | // create), Put with that etag, and if the Put itself loses the race |
| 43 | // (something else wrote in between), re-read once and retry with the fresh |
| 44 | // etag. A second conflict is NOT retried again and is returned to the |
| 45 | // caller: this box is the only writer that should ever be targeting its own |
| 46 | // raw key, so two conflicts in a row means something else collided |
| 47 | // (clock-skewed duplicate machine id, a bug elsewhere) — that deserves a |
| 48 | // loud failure, not a silent third attempt. |
| 49 | func WriteRaw(ctx context.Context, s release.Store, c RawChunk) error { |
| 50 | body, err := json.MarshalIndent(c, "", " ") |
| 51 | if err != nil { |
| 52 | return err |
| 53 | } |
| 54 | key := RawKey(c.Box, c.Hour) |
| 55 | |
| 56 | etag, err := currentRawEtag(ctx, s, key) |
| 57 | if err != nil { |
| 58 | return err |
| 59 | } |
| 60 | if _, err := s.Put(ctx, key, body, etag); err != nil { |
| 61 | if !errors.Is(err, release.ErrCASConflict) { |
| 62 | return err |
| 63 | } |
| 64 | etag, err = currentRawEtag(ctx, s, key) |
| 65 | if err != nil { |
| 66 | return err |
| 67 | } |
| 68 | if _, err := s.Put(ctx, key, body, etag); err != nil { |
| 69 | return err |
| 70 | } |
| 71 | } |
| 72 | return nil |
| 73 | } |
| 74 | |
| 75 | // currentRawEtag reads a key's etag for a replace-only Put, folding |
| 76 | // "doesn't exist yet" into "" (create-only) — the same absent-means-empty |
| 77 | // convention release.Store's own CAS contract uses. |
| 78 | func currentRawEtag(ctx context.Context, s release.Store, key string) (string, error) { |
| 79 | _, etag, err := s.Get(ctx, key) |
| 80 | if err != nil { |
| 81 | if errors.Is(err, os.ErrNotExist) { |
| 82 | return "", nil |
| 83 | } |
| 84 | return "", err |
| 85 | } |
| 86 | return etag, nil |
| 87 | } |
| 88 | |
| 89 | // StaleRawKeys returns the subset of keys (raw chunk keys, as produced by |
| 90 | // RawKey) whose hour is older than now.Add(-keep) — candidates for pruning. |
| 91 | // It is pure: no I/O, and it MUST NEVER return a key it cannot parse — never |
| 92 | // delete what you don't understand. A key is parsed by taking the text |
| 93 | // between the last "/" and ".json" and parsing it with the same UTC hour |
| 94 | // format RawKey/HourOf use; anything that doesn't fit that shape (a |
| 95 | // different key layout, a corrupted or hand-edited name) is silently |
| 96 | // skipped rather than treated as either stale or fresh. |
| 97 | func StaleRawKeys(keys []string, now time.Time, keep time.Duration) []string { |
| 98 | cutoff := now.Add(-keep) |
| 99 | var out []string |
| 100 | for _, key := range keys { |
| 101 | hour, ok := rawKeyHour(key) |
| 102 | if !ok { |
| 103 | continue |
| 104 | } |
| 105 | t, err := time.Parse(hourFormat, hour) |
| 106 | if err != nil { |
| 107 | continue |
| 108 | } |
| 109 | if t.Before(cutoff) { |
| 110 | out = append(out, key) |
| 111 | } |
| 112 | } |
| 113 | return out |
| 114 | } |
| 115 | |
| 116 | // rawKeyHour extracts the "<hour>" portion of a "…/<hour>.json" key. |
| 117 | func rawKeyHour(key string) (string, bool) { |
| 118 | const suffix = ".json" |
| 119 | if !strings.HasSuffix(key, suffix) { |
| 120 | return "", false |
| 121 | } |
| 122 | base := strings.TrimSuffix(key, suffix) |
| 123 | idx := strings.LastIndex(base, "/") |
| 124 | hour := base[idx+1:] |
| 125 | if hour == "" { |
| 126 | return "", false |
| 127 | } |
| 128 | return hour, true |
| 129 | } |
| 130 | |
amadan