@iann29/rastro 0.1.0-alpha.1 → 0.1.0-alpha.11

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (179) hide show
  1. package/README.md +693 -69
  2. package/agent/integration.md +801 -0
  3. package/agent/manifest.json +205 -0
  4. package/agent/manifest.schema.json +444 -0
  5. package/dist/client/federation.d.ts +381 -0
  6. package/dist/client/federation.d.ts.map +1 -0
  7. package/dist/client/federation.js +274 -0
  8. package/dist/client/federation.js.map +1 -0
  9. package/dist/client/index.d.ts +2673 -15
  10. package/dist/client/index.d.ts.map +1 -1
  11. package/dist/client/index.js +565 -33
  12. package/dist/client/index.js.map +1 -1
  13. package/dist/component/_generated/api.d.ts +22 -0
  14. package/dist/component/_generated/api.d.ts.map +1 -1
  15. package/dist/component/_generated/api.js.map +1 -1
  16. package/dist/component/_generated/component.d.ts +315 -4
  17. package/dist/component/_generated/component.d.ts.map +1 -1
  18. package/dist/component/_generated/server.d.ts +4 -0
  19. package/dist/component/_generated/server.d.ts.map +1 -1
  20. package/dist/component/_generated/server.js.map +1 -1
  21. package/dist/component/affiliates.d.ts.map +1 -1
  22. package/dist/component/affiliates.js +6 -2
  23. package/dist/component/affiliates.js.map +1 -1
  24. package/dist/component/cardinality.d.ts +12 -0
  25. package/dist/component/cardinality.d.ts.map +1 -0
  26. package/dist/component/cardinality.js +94 -0
  27. package/dist/component/cardinality.js.map +1 -0
  28. package/dist/component/constants.d.ts +28 -1
  29. package/dist/component/constants.d.ts.map +1 -1
  30. package/dist/component/constants.js +44 -1
  31. package/dist/component/constants.js.map +1 -1
  32. package/dist/component/convex.config.d.ts +6 -1
  33. package/dist/component/convex.config.js +9 -1
  34. package/dist/component/convex.config.js.map +1 -1
  35. package/dist/component/coverage.d.ts +27 -0
  36. package/dist/component/coverage.d.ts.map +1 -0
  37. package/dist/component/coverage.js +56 -0
  38. package/dist/component/coverage.js.map +1 -0
  39. package/dist/component/diagnostics.d.ts +9 -0
  40. package/dist/component/diagnostics.d.ts.map +1 -0
  41. package/dist/component/diagnostics.js +47 -0
  42. package/dist/component/diagnostics.js.map +1 -0
  43. package/dist/component/errors.d.ts +1 -1
  44. package/dist/component/errors.d.ts.map +1 -1
  45. package/dist/component/errors.js.map +1 -1
  46. package/dist/component/eventStore.d.ts +21 -9
  47. package/dist/component/eventStore.d.ts.map +1 -1
  48. package/dist/component/eventStore.js +142 -152
  49. package/dist/component/eventStore.js.map +1 -1
  50. package/dist/component/funnels.d.ts.map +1 -1
  51. package/dist/component/funnels.js +5 -3
  52. package/dist/component/funnels.js.map +1 -1
  53. package/dist/component/geo.d.ts +73 -0
  54. package/dist/component/geo.d.ts.map +1 -0
  55. package/dist/component/geo.js +648 -0
  56. package/dist/component/geo.js.map +1 -0
  57. package/dist/component/goals.d.ts.map +1 -1
  58. package/dist/component/goals.js +8 -5
  59. package/dist/component/goals.js.map +1 -1
  60. package/dist/component/guards.d.ts.map +1 -1
  61. package/dist/component/guards.js.map +1 -1
  62. package/dist/component/http.d.ts.map +1 -1
  63. package/dist/component/http.js +281 -62
  64. package/dist/component/http.js.map +1 -1
  65. package/dist/component/identity.d.ts +13 -0
  66. package/dist/component/identity.d.ts.map +1 -0
  67. package/dist/component/identity.js +58 -0
  68. package/dist/component/identity.js.map +1 -0
  69. package/dist/component/ingest.d.ts +3 -1
  70. package/dist/component/ingest.d.ts.map +1 -1
  71. package/dist/component/ingest.js +498 -95
  72. package/dist/component/ingest.js.map +1 -1
  73. package/dist/component/live.d.ts.map +1 -1
  74. package/dist/component/live.js +7 -6
  75. package/dist/component/live.js.map +1 -1
  76. package/dist/component/localTime.d.ts +25 -0
  77. package/dist/component/localTime.d.ts.map +1 -0
  78. package/dist/component/localTime.js +126 -0
  79. package/dist/component/localTime.js.map +1 -0
  80. package/dist/component/reports.d.ts +266 -10
  81. package/dist/component/reports.d.ts.map +1 -1
  82. package/dist/component/reports.js +1243 -137
  83. package/dist/component/reports.js.map +1 -1
  84. package/dist/component/retention.d.ts +75 -1
  85. package/dist/component/retention.d.ts.map +1 -1
  86. package/dist/component/retention.js +561 -38
  87. package/dist/component/retention.js.map +1 -1
  88. package/dist/component/rollupStore.d.ts +320 -0
  89. package/dist/component/rollupStore.d.ts.map +1 -0
  90. package/dist/component/rollupStore.js +596 -0
  91. package/dist/component/rollupStore.js.map +1 -0
  92. package/dist/component/rollups.d.ts +20 -0
  93. package/dist/component/rollups.d.ts.map +1 -0
  94. package/dist/component/rollups.js +73 -0
  95. package/dist/component/rollups.js.map +1 -0
  96. package/dist/component/sanitize.d.ts +24 -1
  97. package/dist/component/sanitize.d.ts.map +1 -1
  98. package/dist/component/sanitize.js +96 -16
  99. package/dist/component/sanitize.js.map +1 -1
  100. package/dist/component/schema.d.ts +687 -65
  101. package/dist/component/schema.js +187 -20
  102. package/dist/component/schema.js.map +1 -1
  103. package/dist/component/sites.d.ts +12 -0
  104. package/dist/component/sites.d.ts.map +1 -1
  105. package/dist/component/sites.js +41 -7
  106. package/dist/component/sites.js.map +1 -1
  107. package/dist/component/useragent.d.ts +9 -0
  108. package/dist/component/useragent.d.ts.map +1 -0
  109. package/dist/component/useragent.js +152 -0
  110. package/dist/component/useragent.js.map +1 -0
  111. package/dist/component/validators.d.ts +124 -69
  112. package/dist/component/validators.d.ts.map +1 -1
  113. package/dist/component/validators.js +35 -14
  114. package/dist/component/validators.js.map +1 -1
  115. package/dist/component/visitors.d.ts +19 -0
  116. package/dist/component/visitors.d.ts.map +1 -0
  117. package/dist/component/visitors.js +86 -0
  118. package/dist/component/visitors.js.map +1 -0
  119. package/dist/component/vitals.d.ts +41 -0
  120. package/dist/component/vitals.d.ts.map +1 -0
  121. package/dist/component/vitals.js +115 -0
  122. package/dist/component/vitals.js.map +1 -0
  123. package/dist/react/index.d.ts.map +1 -1
  124. package/dist/react/index.js.map +1 -1
  125. package/dist/tracker/generated.d.ts +11 -4
  126. package/dist/tracker/generated.d.ts.map +1 -1
  127. package/dist/tracker/generated.js +11 -4
  128. package/dist/tracker/generated.js.map +1 -1
  129. package/dist/tracker/tracker.d.ts +1 -1
  130. package/dist/tracker/tracker.d.ts.map +1 -1
  131. package/dist/tracker/tracker.js +57 -20
  132. package/dist/tracker/tracker.js.map +1 -1
  133. package/dist/tracker/vitals.d.ts +10 -0
  134. package/dist/tracker/vitals.d.ts.map +1 -0
  135. package/dist/tracker/vitals.js +140 -0
  136. package/dist/tracker/vitals.js.map +1 -0
  137. package/dist/tracker.min.js +1 -1
  138. package/dist/vitals.min.js +1 -0
  139. package/docs/benchmarks/2026-08-20-realistic.md +76 -76
  140. package/docs/benchmarks/2026-08-21-formal-certification.md +353 -0
  141. package/docs/benchmarks/2026-08-30-alpha6-recertification.md +206 -0
  142. package/docs/federation-setup.md +464 -0
  143. package/docs/federation.md +352 -0
  144. package/docs/upgrading.md +344 -0
  145. package/llms.txt +72 -0
  146. package/package.json +55 -11
  147. package/scripts/benchmark-ingest.mjs +175 -73
  148. package/scripts/generate-federation-keys.mjs +20 -0
  149. package/src/component/_generated/api.ts +22 -0
  150. package/src/component/_generated/component.ts +381 -4
  151. package/src/component/_generated/server.ts +4 -0
  152. package/src/component/affiliates.ts +20 -5
  153. package/src/component/cardinality.ts +117 -0
  154. package/src/component/constants.ts +44 -1
  155. package/src/component/convex.config.ts +11 -1
  156. package/src/component/coverage.ts +71 -0
  157. package/src/component/diagnostics.ts +65 -0
  158. package/src/component/errors.ts +2 -1
  159. package/src/component/eventStore.ts +217 -193
  160. package/src/component/funnels.ts +19 -16
  161. package/src/component/geo.ts +835 -0
  162. package/src/component/goals.ts +29 -21
  163. package/src/component/guards.ts +3 -1
  164. package/src/component/http.ts +404 -70
  165. package/src/component/identity.ts +74 -0
  166. package/src/component/ingest.ts +894 -188
  167. package/src/component/live.ts +13 -7
  168. package/src/component/localTime.ts +167 -0
  169. package/src/component/reports.ts +1872 -197
  170. package/src/component/retention.ts +788 -96
  171. package/src/component/rollupStore.ts +799 -0
  172. package/src/component/rollups.ts +82 -0
  173. package/src/component/sanitize.ts +144 -29
  174. package/src/component/schema.ts +217 -21
  175. package/src/component/sites.ts +59 -12
  176. package/src/component/useragent.ts +171 -0
  177. package/src/component/validators.ts +49 -14
  178. package/src/component/visitors.ts +116 -0
  179. package/src/component/vitals.ts +146 -0
@@ -1,12 +1,22 @@
1
1
  import { v } from "convex/values";
2
+ import { internal } from "./_generated/api.js";
2
3
  import type { Id } from "./_generated/dataModel.js";
3
4
  import type { MutationCtx } from "./_generated/server.js";
4
- import { mutation } from "./_generated/server.js";
5
- import { MAX_RETENTION_BATCH } from "./constants.js";
5
+ import { internalMutation, mutation, query } from "./_generated/server.js";
6
+ import {
7
+ DAY_MS,
8
+ HOUR_MS,
9
+ MAX_RETENTION_BATCH,
10
+ MINUTE_MS,
11
+ RATE_LIMIT_WINDOW_MS,
12
+ RETENTION_LEASE_MS,
13
+ RETENTION_POLICY_INTERVAL_MS,
14
+ } from "./constants.js";
15
+ import { ensureAnalyticsControl } from "./coverage.js";
6
16
  import { fail } from "./errors.js";
7
17
  import { deleteTelemetryBefore } from "./eventStore.js";
8
18
 
9
- const cleanupKindValidator = v.union(
19
+ export const cleanupKindValidator = v.union(
10
20
  v.literal("events"),
11
21
  v.literal("sessions"),
12
22
  v.literal("conversions"),
@@ -16,12 +26,23 @@ const cleanupKindValidator = v.union(
16
26
  v.literal("funnelProgress"),
17
27
  v.literal("funnelSteps"),
18
28
  v.literal("affiliateStats"),
29
+ v.literal("goalRollups"),
30
+ v.literal("funnelRollups"),
31
+ v.literal("affiliateRollups"),
32
+ v.literal("vitalRollups"),
33
+ v.literal("siteMapRollups"),
19
34
  v.literal("attributions"),
20
35
  v.literal("liveSessions"),
21
36
  v.literal("ingestWindows"),
22
37
  v.literal("siteIngestWindows"),
23
38
  );
24
39
 
40
+ const cleanupResultValidator = v.object({
41
+ deleted: v.number(),
42
+ hasMore: v.boolean(),
43
+ before: v.number(),
44
+ });
45
+
25
46
  export const cleanup = mutation({
26
47
  args: {
27
48
  siteId: v.id("sites"),
@@ -29,38 +50,385 @@ export const cleanup = mutation({
29
50
  before: v.number(),
30
51
  batchSize: v.optional(v.number()),
31
52
  },
32
- returns: v.object({ deleted: v.number(), hasMore: v.boolean() }),
53
+ returns: cleanupResultValidator,
33
54
  handler: async (ctx, args) => {
34
- if (!Number.isFinite(args.before) || args.before < 0) {
35
- fail("INVALID_ARGUMENT", "before must be a positive timestamp");
55
+ const policy = await ctx.db
56
+ .query("retentionPolicies")
57
+ .withIndex("by_siteId", (range) => range.eq("siteId", args.siteId))
58
+ .unique();
59
+ if (policy) {
60
+ fail(
61
+ "CONFLICT",
62
+ "disable the retention policy before running manual cleanup",
63
+ );
36
64
  }
37
- const batchSize = args.batchSize ?? 100;
38
- if (!Number.isSafeInteger(batchSize) || batchSize < 1 || batchSize > MAX_RETENTION_BATCH) {
39
- fail("INVALID_ARGUMENT", `batchSize must be an integer between 1 and ${MAX_RETENTION_BATCH}`);
65
+ const before = safeCleanupBoundary(args.kind, args.before, Date.now());
66
+ const result = await executeCleanupBatch(ctx, { ...args, before });
67
+ await persistDirectState(ctx, { ...args, before }, result);
68
+ await recordRetentionBoundary(ctx, args.siteId, args.kind, before);
69
+ return { ...result, before };
70
+ },
71
+ });
72
+
73
+ const policyValidator = v.object({
74
+ eventsDays: v.number(),
75
+ sessionsDays: v.number(),
76
+ conversionsDays: v.number(),
77
+ aggregatesDays: v.number(),
78
+ reportsDays: v.number(),
79
+ });
80
+
81
+ export const setPolicy = mutation({
82
+ args: { siteId: v.id("sites"), ...policyValidator.fields },
83
+ returns: v.null(),
84
+ handler: async (ctx, args) => {
85
+ validatePolicy(args);
86
+ const now = Date.now();
87
+ const existingPolicy = await ctx.db
88
+ .query("retentionPolicies")
89
+ .withIndex("by_siteId", (range) => range.eq("siteId", args.siteId))
90
+ .unique();
91
+ const policy = { ...args, updatedAt: now };
92
+ if (existingPolicy) {
93
+ await ctx.db.replace("retentionPolicies", existingPolicy._id, policy);
94
+ } else {
95
+ await ctx.db.insert("retentionPolicies", policy);
40
96
  }
41
- const effectiveBatchSize = args.kind === "events"
42
- ? Math.min(batchSize, 32)
43
- : batchSize;
44
- if (args.kind === "events") {
45
- return await deleteTelemetryBefore(ctx, {
97
+
98
+ for (const kind of POLICY_CLEANUP_KINDS) {
99
+ const requestedBefore = now - retentionDaysForKind(policy, kind) * DAY_MS;
100
+ const before = safeCleanupBoundary(kind, requestedBefore, now);
101
+ const existing = await loadState(ctx, args.siteId, kind);
102
+ const generation = (existing?.generation ?? 0) + 1;
103
+ const fields = {
46
104
  siteId: args.siteId,
47
- before: args.before,
48
- batchSize: effectiveBatchSize,
105
+ kind,
106
+ before,
107
+ status: "pending" as const,
108
+ generation,
109
+ attempt: 0,
110
+ scheduledAt: now,
111
+ deleted: 0,
112
+ updatedAt: now,
113
+ };
114
+ const stateId = existing
115
+ ? (await ctx.db.replace("retentionStates", existing._id, fields),
116
+ existing._id)
117
+ : await ctx.db.insert("retentionStates", fields);
118
+ await ctx.scheduler.runAfter(0, internal.retention.runScheduled, {
119
+ stateId,
120
+ generation,
121
+ });
122
+ }
123
+ return null;
124
+ },
125
+ });
126
+
127
+ export const disablePolicy = mutation({
128
+ args: { siteId: v.id("sites") },
129
+ returns: v.null(),
130
+ handler: async (ctx, args) => {
131
+ const policy = await ctx.db
132
+ .query("retentionPolicies")
133
+ .withIndex("by_siteId", (range) => range.eq("siteId", args.siteId))
134
+ .unique();
135
+ if (policy) await ctx.db.delete("retentionPolicies", policy._id);
136
+ const now = Date.now();
137
+ const states = await ctx.db
138
+ .query("retentionStates")
139
+ .withIndex("by_siteId_and_kind", (range) =>
140
+ range.eq("siteId", args.siteId),
141
+ )
142
+ .take(POLICY_CLEANUP_KINDS.length + 1);
143
+ for (const state of states) {
144
+ await ctx.db.patch("retentionStates", state._id, {
145
+ status: "complete",
146
+ generation: state.generation + 1,
147
+ leaseExpiresAt: undefined,
148
+ scheduledAt: undefined,
149
+ updatedAt: now,
150
+ });
151
+ }
152
+ return null;
153
+ },
154
+ });
155
+
156
+ export const runPolicy = mutation({
157
+ args: {
158
+ siteId: v.id("sites"),
159
+ kind: cleanupKindValidator,
160
+ batchSize: v.optional(v.number()),
161
+ },
162
+ returns: cleanupResultValidator,
163
+ handler: async (ctx, args) => {
164
+ const policy = await ctx.db
165
+ .query("retentionPolicies")
166
+ .withIndex("by_siteId", (range) => range.eq("siteId", args.siteId))
167
+ .unique();
168
+ if (!policy) fail("NOT_FOUND", "retention policy not found");
169
+ const now = Date.now();
170
+ const existing = await loadState(ctx, args.siteId, args.kind);
171
+ if (
172
+ existing?.status === "running" &&
173
+ existing.leaseExpiresAt !== undefined &&
174
+ existing.leaseExpiresAt > now
175
+ ) {
176
+ return { deleted: 0, hasMore: true, before: existing.before };
177
+ }
178
+ const before =
179
+ existing && existing.status !== "complete"
180
+ ? existing.before
181
+ : safeCleanupBoundary(
182
+ args.kind,
183
+ now - retentionDaysForKind(policy, args.kind) * DAY_MS,
184
+ now,
185
+ );
186
+ const result = await executeCleanupBatch(ctx, { ...args, before });
187
+ const generation = existing?.generation ?? 1;
188
+ const fields = {
189
+ siteId: args.siteId,
190
+ kind: args.kind,
191
+ before,
192
+ status: result.hasMore ? ("pending" as const) : ("complete" as const),
193
+ generation,
194
+ attempt: existing?.attempt ?? 0,
195
+ scheduledAt: result.hasMore ? now : now + RETENTION_POLICY_INTERVAL_MS,
196
+ deleted:
197
+ existing?.before === before
198
+ ? existing.deleted + result.deleted
199
+ : result.deleted,
200
+ updatedAt: now,
201
+ };
202
+ const stateId = existing
203
+ ? (await ctx.db.replace("retentionStates", existing._id, fields),
204
+ existing._id)
205
+ : await ctx.db.insert("retentionStates", fields);
206
+ await recordRetentionBoundary(ctx, args.siteId, args.kind, before);
207
+ if (result.hasMore) {
208
+ await ctx.scheduler.runAfter(0, internal.retention.runScheduled, {
209
+ stateId,
210
+ generation,
49
211
  });
212
+ } else {
213
+ await ctx.scheduler.runAt(
214
+ now + RETENTION_POLICY_INTERVAL_MS,
215
+ internal.retention.startScheduledCycle,
216
+ { stateId, generation },
217
+ );
50
218
  }
51
- const deletions = await deletionsForCleanup(
219
+ return { ...result, before };
220
+ },
221
+ });
222
+
223
+ export const runScheduled = internalMutation({
224
+ args: { stateId: v.id("retentionStates"), generation: v.number() },
225
+ returns: v.object({
226
+ claimed: v.boolean(),
227
+ stale: v.boolean(),
228
+ attempt: v.number(),
229
+ }),
230
+ handler: async (ctx, args) => {
231
+ const state = await ctx.db.get("retentionStates", args.stateId);
232
+ const now = Date.now();
233
+ if (
234
+ !state ||
235
+ state.generation !== args.generation ||
236
+ state.status === "complete" ||
237
+ (state.status === "running" &&
238
+ state.leaseExpiresAt !== undefined &&
239
+ state.leaseExpiresAt > now)
240
+ ) {
241
+ return { claimed: false, stale: true, attempt: state?.attempt ?? 0 };
242
+ }
243
+ const attempt = state.attempt + 1;
244
+ const leaseExpiresAt = now + RETENTION_LEASE_MS;
245
+ await ctx.db.patch("retentionStates", state._id, {
246
+ status: "running",
247
+ attempt,
248
+ leaseExpiresAt,
249
+ scheduledAt: now,
250
+ updatedAt: now,
251
+ });
252
+ await ctx.scheduler.runAfter(0, internal.retention.executeScheduled, {
253
+ stateId: state._id,
254
+ generation: state.generation,
255
+ attempt,
256
+ });
257
+ await ctx.scheduler.runAt(leaseExpiresAt, internal.retention.runScheduled, {
258
+ stateId: state._id,
259
+ generation: state.generation,
260
+ });
261
+ return { claimed: true, stale: false, attempt };
262
+ },
263
+ });
264
+
265
+ export const executeScheduled = internalMutation({
266
+ args: {
267
+ stateId: v.id("retentionStates"),
268
+ generation: v.number(),
269
+ attempt: v.number(),
270
+ },
271
+ returns: v.object({
272
+ deleted: v.number(),
273
+ hasMore: v.boolean(),
274
+ stale: v.boolean(),
275
+ }),
276
+ handler: async (ctx, args) => {
277
+ const state = await ctx.db.get("retentionStates", args.stateId);
278
+ if (
279
+ !state ||
280
+ state.generation !== args.generation ||
281
+ state.attempt !== args.attempt ||
282
+ state.status !== "running"
283
+ ) {
284
+ return { deleted: 0, hasMore: false, stale: true };
285
+ }
286
+ const result = await executeCleanupBatch(ctx, {
287
+ siteId: state.siteId,
288
+ kind: state.kind as CleanupKind,
289
+ before: state.before,
290
+ batchSize: 100,
291
+ });
292
+ const now = Date.now();
293
+ const policy = result.hasMore
294
+ ? null
295
+ : await ctx.db
296
+ .query("retentionPolicies")
297
+ .withIndex("by_siteId", (range) => range.eq("siteId", state.siteId))
298
+ .unique();
299
+ const nextCycleAt = policy ? now + RETENTION_POLICY_INTERVAL_MS : undefined;
300
+ await ctx.db.patch("retentionStates", state._id, {
301
+ status: result.hasMore ? "pending" : "complete",
302
+ leaseExpiresAt: undefined,
303
+ scheduledAt: result.hasMore ? now : nextCycleAt,
304
+ deleted: state.deleted + result.deleted,
305
+ updatedAt: now,
306
+ });
307
+ await recordRetentionBoundary(
52
308
  ctx,
53
- args.siteId,
54
- args.kind,
55
- args.before,
56
- effectiveBatchSize + 1,
309
+ state.siteId,
310
+ state.kind as CleanupKind,
311
+ state.before,
57
312
  );
58
- for (const remove of deletions.slice(0, effectiveBatchSize)) {
59
- await remove();
313
+ if (result.hasMore) {
314
+ await ctx.scheduler.runAfter(0, internal.retention.runScheduled, {
315
+ stateId: state._id,
316
+ generation: state.generation,
317
+ });
318
+ } else if (nextCycleAt !== undefined) {
319
+ await ctx.scheduler.runAt(
320
+ nextCycleAt,
321
+ internal.retention.startScheduledCycle,
322
+ { stateId: state._id, generation: state.generation },
323
+ );
60
324
  }
325
+ return { ...result, stale: false };
326
+ },
327
+ });
328
+
329
+ export const startScheduledCycle = internalMutation({
330
+ args: { stateId: v.id("retentionStates"), generation: v.number() },
331
+ returns: v.object({
332
+ started: v.boolean(),
333
+ stale: v.boolean(),
334
+ before: v.union(v.number(), v.null()),
335
+ }),
336
+ handler: async (ctx, args) => {
337
+ const state = await ctx.db.get("retentionStates", args.stateId);
338
+ const now = Date.now();
339
+ if (
340
+ !state ||
341
+ state.generation !== args.generation ||
342
+ state.status !== "complete" ||
343
+ (state.scheduledAt !== undefined && state.scheduledAt > now)
344
+ ) {
345
+ return { started: false, stale: true, before: null };
346
+ }
347
+ const policy = await ctx.db
348
+ .query("retentionPolicies")
349
+ .withIndex("by_siteId", (range) => range.eq("siteId", state.siteId))
350
+ .unique();
351
+ if (!policy) return { started: false, stale: true, before: null };
352
+ const kind = state.kind as CleanupKind;
353
+ const before = safeCleanupBoundary(
354
+ kind,
355
+ now - retentionDaysForKind(policy, kind) * DAY_MS,
356
+ now,
357
+ );
358
+ await ctx.db.patch("retentionStates", state._id, {
359
+ before,
360
+ status: "pending",
361
+ attempt: 0,
362
+ leaseExpiresAt: undefined,
363
+ scheduledAt: now,
364
+ deleted: 0,
365
+ updatedAt: now,
366
+ });
367
+ await ctx.scheduler.runAfter(0, internal.retention.runScheduled, {
368
+ stateId: state._id,
369
+ generation: state.generation,
370
+ });
371
+ return { started: true, stale: false, before };
372
+ },
373
+ });
374
+
375
+ export const status = query({
376
+ args: { siteId: v.id("sites") },
377
+ returns: v.object({
378
+ policy: v.union(
379
+ policyValidator.extend({ updatedAt: v.number() }),
380
+ v.null(),
381
+ ),
382
+ jobs: v.array(
383
+ v.object({
384
+ kind: cleanupKindValidator,
385
+ before: v.number(),
386
+ status: v.union(
387
+ v.literal("pending"),
388
+ v.literal("running"),
389
+ v.literal("complete"),
390
+ ),
391
+ deleted: v.number(),
392
+ scheduledAt: v.union(v.number(), v.null()),
393
+ updatedAt: v.number(),
394
+ }),
395
+ ),
396
+ backlog: v.number(),
397
+ }),
398
+ handler: async (ctx, args) => {
399
+ const policy = await ctx.db
400
+ .query("retentionPolicies")
401
+ .withIndex("by_siteId", (range) => range.eq("siteId", args.siteId))
402
+ .unique();
403
+ const jobs = await ctx.db
404
+ .query("retentionStates")
405
+ .withIndex("by_siteId_and_kind", (range) =>
406
+ range.eq("siteId", args.siteId),
407
+ )
408
+ .take(POLICY_CLEANUP_KINDS.length + 1);
409
+ const publicJobs = jobs.map(
410
+ ({ kind, before, status, deleted, scheduledAt, updatedAt }) => ({
411
+ kind: kind as CleanupKind,
412
+ before,
413
+ status,
414
+ deleted,
415
+ scheduledAt: scheduledAt ?? null,
416
+ updatedAt,
417
+ }),
418
+ );
61
419
  return {
62
- deleted: Math.min(deletions.length, effectiveBatchSize),
63
- hasMore: deletions.length > effectiveBatchSize,
420
+ policy: policy
421
+ ? {
422
+ eventsDays: policy.eventsDays,
423
+ sessionsDays: policy.sessionsDays,
424
+ conversionsDays: policy.conversionsDays,
425
+ aggregatesDays: policy.aggregatesDays,
426
+ reportsDays: policy.reportsDays,
427
+ updatedAt: policy.updatedAt,
428
+ }
429
+ : null,
430
+ jobs: publicJobs,
431
+ backlog: publicJobs.filter((job) => job.status !== "complete").length,
64
432
  };
65
433
  },
66
434
  });
@@ -75,11 +443,247 @@ type CleanupKind =
75
443
  | "funnelProgress"
76
444
  | "funnelSteps"
77
445
  | "affiliateStats"
446
+ | "goalRollups"
447
+ | "funnelRollups"
448
+ | "affiliateRollups"
449
+ | "vitalRollups"
450
+ | "siteMapRollups"
78
451
  | "attributions"
79
452
  | "liveSessions"
80
453
  | "ingestWindows"
81
454
  | "siteIngestWindows";
82
455
 
456
+ const POLICY_CLEANUP_KINDS: CleanupKind[] = [
457
+ "events",
458
+ "sessions",
459
+ "conversions",
460
+ "aggregateHour",
461
+ "aggregateDay",
462
+ "goalCompletions",
463
+ "funnelProgress",
464
+ "funnelSteps",
465
+ "affiliateStats",
466
+ "goalRollups",
467
+ "funnelRollups",
468
+ "affiliateRollups",
469
+ "vitalRollups",
470
+ "siteMapRollups",
471
+ "attributions",
472
+ "ingestWindows",
473
+ "siteIngestWindows",
474
+ ];
475
+
476
+ function validatePolicy(policy: {
477
+ eventsDays: number;
478
+ sessionsDays: number;
479
+ conversionsDays: number;
480
+ aggregatesDays: number;
481
+ reportsDays: number;
482
+ }) {
483
+ for (const [name, days] of Object.entries(policy)) {
484
+ if (!name.endsWith("Days")) continue;
485
+ if (!Number.isSafeInteger(days) || days < 1 || days > 3_650) {
486
+ fail("INVALID_ARGUMENT", `${name} must be an integer between 1 and 3650`);
487
+ }
488
+ }
489
+ }
490
+
491
+ function retentionDaysForKind(
492
+ policy: {
493
+ eventsDays: number;
494
+ sessionsDays: number;
495
+ conversionsDays: number;
496
+ aggregatesDays: number;
497
+ reportsDays: number;
498
+ },
499
+ kind: CleanupKind,
500
+ ) {
501
+ if (kind === "events") return policy.eventsDays;
502
+ if (kind === "sessions") return policy.sessionsDays;
503
+ if (kind === "conversions") return policy.conversionsDays;
504
+ if (kind === "aggregateHour" || kind === "aggregateDay") {
505
+ return policy.aggregatesDays;
506
+ }
507
+ return policy.reportsDays;
508
+ }
509
+
510
+ function safeCleanupBoundary(
511
+ kind: CleanupKind,
512
+ requested: number,
513
+ now: number,
514
+ ) {
515
+ if (!Number.isFinite(requested) || requested < 0) {
516
+ fail("INVALID_ARGUMENT", "before must be a positive timestamp");
517
+ }
518
+ if (kind === "events") return safeBucketBoundary(requested, now, MINUTE_MS);
519
+ if (kind === "aggregateHour")
520
+ return safeBucketBoundary(requested, now, HOUR_MS);
521
+ if (
522
+ kind === "aggregateDay" ||
523
+ kind === "funnelSteps" ||
524
+ kind === "affiliateStats" ||
525
+ kind === "goalRollups" ||
526
+ kind === "funnelRollups" ||
527
+ kind === "affiliateRollups" ||
528
+ kind === "vitalRollups" ||
529
+ kind === "siteMapRollups"
530
+ ) {
531
+ return safeBucketBoundary(requested, now, DAY_MS);
532
+ }
533
+ if (kind === "ingestWindows" || kind === "siteIngestWindows") {
534
+ return Math.min(
535
+ Math.floor(requested / RATE_LIMIT_WINDOW_MS) * RATE_LIMIT_WINDOW_MS,
536
+ Math.floor(now / RATE_LIMIT_WINDOW_MS) * RATE_LIMIT_WINDOW_MS,
537
+ );
538
+ }
539
+ return Math.trunc(requested);
540
+ }
541
+
542
+ function safeBucketBoundary(
543
+ requested: number,
544
+ now: number,
545
+ intervalMs: number,
546
+ ) {
547
+ return Math.min(
548
+ Math.floor(requested / intervalMs) * intervalMs,
549
+ Math.floor(now / intervalMs) * intervalMs,
550
+ );
551
+ }
552
+
553
+ async function loadState(
554
+ ctx: MutationCtx,
555
+ siteId: Id<"sites">,
556
+ kind: CleanupKind,
557
+ ) {
558
+ return ctx.db
559
+ .query("retentionStates")
560
+ .withIndex("by_siteId_and_kind", (range) =>
561
+ range.eq("siteId", siteId).eq("kind", kind),
562
+ )
563
+ .unique();
564
+ }
565
+
566
+ async function persistDirectState(
567
+ ctx: MutationCtx,
568
+ args: { siteId: Id<"sites">; kind: CleanupKind; before: number },
569
+ result: { deleted: number; hasMore: boolean },
570
+ ) {
571
+ const now = Date.now();
572
+ const existing = await loadState(ctx, args.siteId, args.kind);
573
+ const fields = {
574
+ siteId: args.siteId,
575
+ kind: args.kind,
576
+ before: args.before,
577
+ status: result.hasMore ? ("pending" as const) : ("complete" as const),
578
+ generation: (existing?.generation ?? 0) + 1,
579
+ attempt: 0,
580
+ deleted:
581
+ existing?.before === args.before
582
+ ? existing.deleted + result.deleted
583
+ : result.deleted,
584
+ updatedAt: now,
585
+ };
586
+ if (existing) {
587
+ await ctx.db.replace("retentionStates", existing._id, fields);
588
+ } else {
589
+ await ctx.db.insert("retentionStates", fields);
590
+ }
591
+ }
592
+
593
+ async function recordRetentionBoundary(
594
+ ctx: MutationCtx,
595
+ siteId: Id<"sites">,
596
+ kind: CleanupKind,
597
+ before: number,
598
+ ) {
599
+ const datasets = coverageDatasetsForKind(kind);
600
+ if (datasets.length === 0) return;
601
+ const now = Date.now();
602
+ const coverage = await ensureAnalyticsControl(
603
+ ctx,
604
+ siteId,
605
+ now,
606
+ datasets.map((dataset) => ({ dataset, before })),
607
+ );
608
+ const retained = coverage.retained.filter(
609
+ (item) => !datasets.includes(item.dataset),
610
+ );
611
+ let changed = retained.length !== coverage.retained.length;
612
+ for (const dataset of datasets) {
613
+ const prior =
614
+ coverage.retained.find((item) => item.dataset === dataset)?.before ?? 0;
615
+ const next = Math.max(prior, before);
616
+ retained.push({ dataset, before: next });
617
+ changed ||= next !== prior;
618
+ }
619
+ if (changed) {
620
+ await ctx.db.patch("analyticsCoverage", coverage._id, {
621
+ retained,
622
+ updatedAt: now,
623
+ });
624
+ }
625
+ }
626
+
627
+ function coverageDatasetsForKind(kind: CleanupKind): string[] {
628
+ if (kind === "events") return ["events"];
629
+ if (kind === "sessions") return ["sessions"];
630
+ if (kind === "conversions") return ["conversions"];
631
+ if (kind === "aggregateHour") return ["overviewHour"];
632
+ if (kind === "aggregateDay") return ["overviewDay", "overviewLocalDay"];
633
+ if (kind === "goalCompletions") return ["goalCompletions"];
634
+ if (kind === "funnelSteps") return ["funnelSteps"];
635
+ if (kind === "affiliateStats") return ["affiliateStats"];
636
+ if (kind === "goalRollups") return ["goalRollups"];
637
+ if (kind === "funnelRollups") return ["funnelRollups"];
638
+ if (kind === "affiliateRollups") return ["affiliateRollups"];
639
+ if (kind === "vitalRollups") return ["vitalRollups"];
640
+ if (kind === "siteMapRollups") return ["siteMapRollups"];
641
+ return [];
642
+ }
643
+
644
+ async function executeCleanupBatch(
645
+ ctx: MutationCtx,
646
+ args: {
647
+ siteId: Id<"sites">;
648
+ kind: CleanupKind;
649
+ before: number;
650
+ batchSize?: number;
651
+ },
652
+ ) {
653
+ const batchSize = args.batchSize ?? 100;
654
+ if (
655
+ !Number.isSafeInteger(batchSize) ||
656
+ batchSize < 1 ||
657
+ batchSize > MAX_RETENTION_BATCH
658
+ ) {
659
+ fail(
660
+ "INVALID_ARGUMENT",
661
+ `batchSize must be an integer between 1 and ${MAX_RETENTION_BATCH}`,
662
+ );
663
+ }
664
+ const effectiveBatchSize =
665
+ args.kind === "events" ? Math.min(batchSize, 32) : batchSize;
666
+ if (args.kind === "events") {
667
+ return deleteTelemetryBefore(ctx, {
668
+ siteId: args.siteId,
669
+ before: args.before,
670
+ batchSize: effectiveBatchSize,
671
+ });
672
+ }
673
+ const deletions = await deletionsForCleanup(
674
+ ctx,
675
+ args.siteId,
676
+ args.kind,
677
+ args.before,
678
+ effectiveBatchSize + 1,
679
+ );
680
+ for (const remove of deletions.slice(0, effectiveBatchSize)) await remove();
681
+ return {
682
+ deleted: Math.min(deletions.length, effectiveBatchSize),
683
+ hasMore: deletions.length > effectiveBatchSize,
684
+ };
685
+ }
686
+
83
687
  type CleanupDelete = () => Promise<void>;
84
688
 
85
689
  async function deletionsForCleanup(
@@ -90,94 +694,182 @@ async function deletionsForCleanup(
90
694
  limit: number,
91
695
  ): Promise<CleanupDelete[]> {
92
696
  if (kind === "sessions") {
93
- return (await ctx.db
94
- .query("sessions")
95
- .withIndex("by_siteId_and_lastSeenAt", (range) =>
96
- range.eq("siteId", siteId).lt("lastSeenAt", before),
97
- )
98
- .take(limit)).map((row) => () => ctx.db.delete("sessions", row._id));
697
+ return (
698
+ await ctx.db
699
+ .query("sessions")
700
+ .withIndex("by_siteId_and_lastSeenAt", (q) =>
701
+ q.eq("siteId", siteId).lt("lastSeenAt", before),
702
+ )
703
+ .take(limit)
704
+ ).map((row) => () => ctx.db.delete("sessions", row._id));
99
705
  }
100
706
  if (kind === "conversions") {
101
- return (await ctx.db
102
- .query("conversions")
103
- .withIndex("by_siteId_and_timestamp", (range) =>
104
- range.eq("siteId", siteId).lt("timestamp", before),
105
- )
106
- .take(limit)).map((row) => () => ctx.db.delete("conversions", row._id));
707
+ return (
708
+ await ctx.db
709
+ .query("conversions")
710
+ .withIndex("by_siteId_and_timestamp", (q) =>
711
+ q.eq("siteId", siteId).lt("timestamp", before),
712
+ )
713
+ .take(limit)
714
+ ).map((row) => () => ctx.db.delete("conversions", row._id));
107
715
  }
108
716
  if (kind === "goalCompletions") {
109
- return (await ctx.db
110
- .query("goalCompletions")
111
- .withIndex("by_siteId_and_timestamp", (range) =>
112
- range.eq("siteId", siteId).lt("timestamp", before),
113
- )
114
- .take(limit)).map((row) => () => ctx.db.delete("goalCompletions", row._id));
717
+ return (
718
+ await ctx.db
719
+ .query("goalCompletions")
720
+ .withIndex("by_siteId_and_timestamp", (q) =>
721
+ q.eq("siteId", siteId).lt("timestamp", before),
722
+ )
723
+ .take(limit)
724
+ ).map((row) => () => ctx.db.delete("goalCompletions", row._id));
115
725
  }
116
726
  if (kind === "aggregateHour" || kind === "aggregateDay") {
117
- return (await ctx.db
118
- .query("aggregateBuckets")
119
- .withIndex("by_siteId_and_granularity_and_bucketStart", (range) =>
120
- range
121
- .eq("siteId", siteId)
122
- .eq("granularity", kind === "aggregateHour" ? "hour" : "day")
123
- .lt("bucketStart", before),
124
- )
125
- .take(limit)).map((row) => () => ctx.db.delete("aggregateBuckets", row._id));
727
+ // The daily kind also retires the site's local-day buckets. Their cutoff
728
+ // sits a day and an hour earlier than the UTC-aligned one: a local day
729
+ // may begin almost a day before its UTC counterpart and, across a
730
+ // fall-back transition, last 25 hours, so this keeps every bucket that
731
+ // could still contain the cutoff instant.
732
+ const granularities =
733
+ kind === "aggregateHour"
734
+ ? ([["hour", before]] as const)
735
+ : ([
736
+ ["day", before],
737
+ ["localDay", before - DAY_MS - HOUR_MS],
738
+ ] as const);
739
+ const deletions: Array<() => Promise<void>> = [];
740
+ for (const [granularity, cutoff] of granularities) {
741
+ if (deletions.length >= limit) break;
742
+ const rows = await ctx.db
743
+ .query("aggregateBuckets")
744
+ .withIndex("by_siteId_and_granularity_and_bucketStart", (q) =>
745
+ q
746
+ .eq("siteId", siteId)
747
+ .eq("granularity", granularity)
748
+ .lt("bucketStart", cutoff),
749
+ )
750
+ .take(limit - deletions.length);
751
+ deletions.push(
752
+ ...rows.map((row) => () => ctx.db.delete("aggregateBuckets", row._id)),
753
+ );
754
+ }
755
+ return deletions;
126
756
  }
127
757
  if (kind === "funnelProgress") {
128
- return (await ctx.db
129
- .query("funnelProgress")
130
- .withIndex("by_siteId_and_lastStepAt", (range) =>
131
- range.eq("siteId", siteId).lt("lastStepAt", before),
132
- )
133
- .take(limit)).map((row) => () => ctx.db.delete("funnelProgress", row._id));
758
+ return (
759
+ await ctx.db
760
+ .query("funnelProgress")
761
+ .withIndex("by_siteId_and_lastStepAt", (q) =>
762
+ q.eq("siteId", siteId).lt("lastStepAt", before),
763
+ )
764
+ .take(limit)
765
+ ).map((row) => () => ctx.db.delete("funnelProgress", row._id));
134
766
  }
135
767
  if (kind === "funnelSteps") {
136
- return (await ctx.db
137
- .query("funnelStepBuckets")
138
- .withIndex("by_siteId_and_bucketStart", (range) =>
139
- range.eq("siteId", siteId).lt("bucketStart", before),
140
- )
141
- .take(limit)).map((row) => () => ctx.db.delete("funnelStepBuckets", row._id));
768
+ return (
769
+ await ctx.db
770
+ .query("funnelStepBuckets")
771
+ .withIndex("by_siteId_and_bucketStart", (q) =>
772
+ q.eq("siteId", siteId).lt("bucketStart", before),
773
+ )
774
+ .take(limit)
775
+ ).map((row) => () => ctx.db.delete("funnelStepBuckets", row._id));
142
776
  }
143
777
  if (kind === "affiliateStats") {
144
- return (await ctx.db
145
- .query("affiliateDailyStats")
146
- .withIndex("by_siteId_and_bucketStart", (range) =>
147
- range.eq("siteId", siteId).lt("bucketStart", before),
148
- )
149
- .take(limit)).map((row) => () => ctx.db.delete("affiliateDailyStats", row._id));
778
+ return (
779
+ await ctx.db
780
+ .query("affiliateDailyStats")
781
+ .withIndex("by_siteId_and_bucketStart", (q) =>
782
+ q.eq("siteId", siteId).lt("bucketStart", before),
783
+ )
784
+ .take(limit)
785
+ ).map((row) => () => ctx.db.delete("affiliateDailyStats", row._id));
786
+ }
787
+ if (kind === "goalRollups") {
788
+ return (
789
+ await ctx.db
790
+ .query("goalDailyRollups")
791
+ .withIndex("by_siteId_and_bucketStart", (q) =>
792
+ q.eq("siteId", siteId).lt("bucketStart", before),
793
+ )
794
+ .take(limit)
795
+ ).map((row) => () => ctx.db.delete("goalDailyRollups", row._id));
796
+ }
797
+ if (kind === "funnelRollups") {
798
+ return (
799
+ await ctx.db
800
+ .query("funnelDailyRollups")
801
+ .withIndex("by_siteId_and_bucketStart", (q) =>
802
+ q.eq("siteId", siteId).lt("bucketStart", before),
803
+ )
804
+ .take(limit)
805
+ ).map((row) => () => ctx.db.delete("funnelDailyRollups", row._id));
806
+ }
807
+ if (kind === "affiliateRollups") {
808
+ return (
809
+ await ctx.db
810
+ .query("affiliateDailyRollups")
811
+ .withIndex("by_siteId_and_bucketStart", (q) =>
812
+ q.eq("siteId", siteId).lt("bucketStart", before),
813
+ )
814
+ .take(limit)
815
+ ).map((row) => () => ctx.db.delete("affiliateDailyRollups", row._id));
816
+ }
817
+ if (kind === "vitalRollups") {
818
+ return (
819
+ await ctx.db
820
+ .query("vitalRollups")
821
+ .withIndex("by_siteId_and_bucketStart", (q) =>
822
+ q.eq("siteId", siteId).lt("bucketStart", before),
823
+ )
824
+ .take(limit)
825
+ ).map((row) => () => ctx.db.delete("vitalRollups", row._id));
826
+ }
827
+ if (kind === "siteMapRollups") {
828
+ return (
829
+ await ctx.db
830
+ .query("routeDailyRollups")
831
+ .withIndex("by_siteId_and_bucketStart", (q) =>
832
+ q.eq("siteId", siteId).lt("bucketStart", before),
833
+ )
834
+ .take(limit)
835
+ ).map((row) => () => ctx.db.delete("routeDailyRollups", row._id));
150
836
  }
151
837
  if (kind === "attributions") {
152
- return (await ctx.db
153
- .query("visitorAttributions")
154
- .withIndex("by_siteId_and_expiresAt", (range) =>
155
- range.eq("siteId", siteId).lt("expiresAt", before),
156
- )
157
- .take(limit)).map((row) => () =>
158
- ctx.db.delete("visitorAttributions", row._id),
159
- );
838
+ return (
839
+ await ctx.db
840
+ .query("visitorAttributions")
841
+ .withIndex("by_siteId_and_expiresAt", (q) =>
842
+ q.eq("siteId", siteId).lt("expiresAt", before),
843
+ )
844
+ .take(limit)
845
+ ).map((row) => () => ctx.db.delete("visitorAttributions", row._id));
160
846
  }
161
847
  if (kind === "liveSessions") {
162
- return (await ctx.db
163
- .query("liveSessions")
164
- .withIndex("by_siteId_and_expiresAt", (range) =>
165
- range.eq("siteId", siteId).lt("expiresAt", before),
166
- )
167
- .take(limit)).map((row) => () => ctx.db.delete("liveSessions", row._id));
848
+ return (
849
+ await ctx.db
850
+ .query("liveSessions")
851
+ .withIndex("by_siteId_and_expiresAt", (q) =>
852
+ q.eq("siteId", siteId).lt("expiresAt", before),
853
+ )
854
+ .take(limit)
855
+ ).map((row) => () => ctx.db.delete("liveSessions", row._id));
168
856
  }
169
857
  if (kind === "siteIngestWindows") {
170
- return (await ctx.db
171
- .query("siteIngestWindows")
172
- .withIndex("by_siteId_and_updatedAt", (range) =>
173
- range.eq("siteId", siteId).lt("updatedAt", before),
174
- )
175
- .take(limit)).map((row) => () => ctx.db.delete("siteIngestWindows", row._id));
858
+ return (
859
+ await ctx.db
860
+ .query("siteIngestWindows")
861
+ .withIndex("by_siteId_and_updatedAt", (q) =>
862
+ q.eq("siteId", siteId).lt("updatedAt", before),
863
+ )
864
+ .take(limit)
865
+ ).map((row) => () => ctx.db.delete("siteIngestWindows", row._id));
176
866
  }
177
- return (await ctx.db
178
- .query("ingestWindows")
179
- .withIndex("by_siteId_and_updatedAt", (range) =>
180
- range.eq("siteId", siteId).lt("updatedAt", before),
181
- )
182
- .take(limit)).map((row) => () => ctx.db.delete("ingestWindows", row._id));
867
+ return (
868
+ await ctx.db
869
+ .query("ingestWindows")
870
+ .withIndex("by_siteId_and_updatedAt", (q) =>
871
+ q.eq("siteId", siteId).lt("updatedAt", before),
872
+ )
873
+ .take(limit)
874
+ ).map((row) => () => ctx.db.delete("ingestWindows", row._id));
183
875
  }