test(imaging): retry transient edge failures instead of flaking the suite
CI / check (push) Successful in 37s
CI / tests-integration (push) Successful in 2m3s
CI / tests-unit (push) Successful in 2m8s
CI / tests-imaging-browsers (push) Successful in 2m15s
CI / tests-ui (push) Successful in 2m54s
CI / preflight (push) Skipped
CI / tests-quality (push) Successful in 1m14s
CI / deploy (push) Successful in 3m37s
CI / check (push) Successful in 37s
CI / tests-integration (push) Successful in 2m3s
CI / tests-unit (push) Successful in 2m8s
CI / tests-imaging-browsers (push) Successful in 2m15s
CI / tests-ui (push) Successful in 2m54s
CI / preflight (push) Skipped
CI / tests-quality (push) Successful in 1m14s
CI / deploy (push) Successful in 3m37s
The cross-browser imaging job runs against the public origin, so every request goes through Cloudflare. One chromium run of "rejects malformed figures › missing figure" came back 525 (Cloudflare: TLS handshake to the origin failed) after a 40s stall — the request never reached the CMS, so a 400 was impossible to observe and the assertion was effectively testing the network. It passed on both firefox and webkit, which is the signature of an edge hiccup rather than a regression. `e2e/imaging/imaging-target.ts` now owns `getThroughEdge` / `fetchThroughEdge`, which retry on the statuses that mean "the app was never reached" (502/503/504 plus Cloudflare's 520-527) with a short backoff, and both edge-fronted specs use them. A genuinely broken edge still exhausts the retries and fails the run, so nothing is masked. Verified: 84 passed / 21 skipped / 0 failed against epicnabbo.nl with IMAGING_EDGE=1 across chromium, firefox and webkit.
This commit is contained in:
1 parent
77a0d1010c
commit
480343d5a5
3 files changed
+117
-22
No files matched your search
@@ -1,3 +1,5 @@
|
||||
import type { APIRequestContext, APIResponse } from "@playwright/test";
|
||||
|
||||
/**
|
||||
* Where the imaging specs are pointed.
|
||||
*
|
||||
@@ -12,6 +14,56 @@
|
||||
*/
|
||||
export const IS_EDGE = process.env.IMAGING_EDGE === "1";
|
||||
|
||||
/**
|
||||
* Failures that mean "the request never reached the app".
|
||||
*
|
||||
* Cloudflare answers the 520-527 family when its own handshake to the origin
|
||||
* fails, and gateways add 502/503/504. Retrying those is not papering over a
|
||||
* bug: the assertion under test is about the CMS, and a 525 says nothing about
|
||||
* it. A genuinely broken edge still exhausts the retries and fails the test.
|
||||
*/
|
||||
const TRANSIENT_EDGE_STATUSES = new Set([
|
||||
502, 503, 504, 520, 521, 522, 523, 524, 525, 526, 527,
|
||||
]);
|
||||
const EDGE_RETRY_ATTEMPTS = 3;
|
||||
const EDGE_RETRY_DELAY_MS = 500;
|
||||
|
||||
async function withEdgeRetry(
|
||||
send: () => Promise<APIResponse>,
|
||||
): Promise<APIResponse> {
|
||||
let response = await send();
|
||||
for (
|
||||
let attempt = 1;
|
||||
attempt < EDGE_RETRY_ATTEMPTS &&
|
||||
TRANSIENT_EDGE_STATUSES.has(response.status());
|
||||
attempt += 1
|
||||
) {
|
||||
await new Promise((resolve) =>
|
||||
setTimeout(resolve, EDGE_RETRY_DELAY_MS * attempt),
|
||||
);
|
||||
response = await send();
|
||||
}
|
||||
return response;
|
||||
}
|
||||
|
||||
/** GET that survives a transient edge/transport failure. */
|
||||
export function getThroughEdge(
|
||||
request: APIRequestContext,
|
||||
url: string,
|
||||
options: Parameters<APIRequestContext["get"]>[1] = {},
|
||||
): Promise<APIResponse> {
|
||||
return withEdgeRetry(() => request.get(url, options));
|
||||
}
|
||||
|
||||
/** POST/OPTIONS/… counterpart of `getThroughEdge`. */
|
||||
export function fetchThroughEdge(
|
||||
request: APIRequestContext,
|
||||
url: string,
|
||||
options: Parameters<APIRequestContext["fetch"]>[1] = {},
|
||||
): Promise<APIResponse> {
|
||||
return withEdgeRetry(() => request.fetch(url, options));
|
||||
}
|
||||
|
||||
// The exact figure that regressed in production: the game client appends a
|
||||
// trailing `.<n>` segment that the renderer ignores, but the previous
|
||||
// validation regex rejected with a 400.
|
||||
|
||||
Reference in new issue
Block a user