feat(observability): report silent catalog UX degradations to Sentry
The catalog failures that hurt UX — a drill/parts fetch that fails into an empty
"couldn't load" panel, or a decoded vehicle whose category tree comes back empty
("model var ama parça yok") — all return HTTP 200 with a degraded body. Nothing
throws, so the global Sentry exception filter never sees them and they go
unnoticed (serkan's complaint was exactly this class). Report them explicitly.
- new common/catalog-degradation.ts: reportCatalogDegradation(kind, ctx),
fingerprinted by kind+source+brand so each failure mode collapses into one
countable Sentry issue (e.g. "drill-load-error · pl24/Ford — N events, M users").
- categories.service: capture on getCategoryWithParts loadError and on an empty
getCategoryTree, Redis-deduped to <=1 event/hour per category/vehicle so a
broken catalog can't flood the stream; telemetry never throws into the request.
tsc + biome clean, categories suite 10/10.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
58
apps/api/src/common/catalog-degradation.ts
Normal file
58
apps/api/src/common/catalog-degradation.ts
Normal file
@@ -0,0 +1,58 @@
|
||||
import * as Sentry from "@sentry/nestjs";
|
||||
|
||||
/**
|
||||
* A *silent* catalog UX failure: the request succeeds (HTTP 200) but the user
|
||||
* gets a degraded result — an empty parts panel / retryable load error, or an
|
||||
* empty category tree on a vehicle that decoded fine. Because nothing throws,
|
||||
* the global exception filter never sees these, so they go unnoticed — exactly
|
||||
* how serkan's "şase girdim, model var ama parça yok" sat invisible. We report
|
||||
* them to Sentry explicitly so they surface and get triaged like real errors.
|
||||
*/
|
||||
export type CatalogDegradationKind =
|
||||
| "empty-tree" // vehicle decoded but the category tree came back empty
|
||||
| "drill-load-error"; // a category drill/parts fetch failed → empty/retry panel, not parts
|
||||
|
||||
export interface CatalogDegradationContext {
|
||||
vehicleId?: string | null;
|
||||
vin?: string | null;
|
||||
brand?: string | null;
|
||||
model?: string | null;
|
||||
source?: string | null;
|
||||
categoryId?: string | null;
|
||||
categoryName?: string | null;
|
||||
linkPath?: string | null;
|
||||
}
|
||||
|
||||
/**
|
||||
* Send a catalog degradation to Sentry as a warning. Fingerprinted by
|
||||
* kind + source + brand so every instance of one failure mode collapses into a
|
||||
* single, countable Sentry issue ("drill-load-error · pl24/Ford — 1.2k events,
|
||||
* 80 users") rather than thousands of unique events. Call sites must dedup
|
||||
* (e.g. a short Redis TTL per vehicle/category) before invoking this, and must
|
||||
* never let it throw into the request path.
|
||||
*/
|
||||
export function reportCatalogDegradation(
|
||||
kind: CatalogDegradationKind,
|
||||
ctx: CatalogDegradationContext,
|
||||
): void {
|
||||
const source = ctx.source ?? "unknown";
|
||||
const brand = ctx.brand ?? "unknown";
|
||||
Sentry.captureMessage(`catalog degraded: ${kind} (${source}/${brand})`, {
|
||||
level: "warning",
|
||||
tags: {
|
||||
catalog_degradation: kind,
|
||||
catalog_source: source,
|
||||
catalog_brand: brand,
|
||||
},
|
||||
// Group by failure mode, not by individual vehicle/category.
|
||||
fingerprint: ["catalog-degradation", kind, source, brand.toLowerCase()],
|
||||
extra: {
|
||||
vehicleId: ctx.vehicleId ?? null,
|
||||
vin: ctx.vin ?? null,
|
||||
model: ctx.model ?? null,
|
||||
categoryId: ctx.categoryId ?? null,
|
||||
categoryName: ctx.categoryName ?? null,
|
||||
linkPath: ctx.linkPath ?? null,
|
||||
},
|
||||
});
|
||||
}
|
||||
Reference in New Issue
Block a user