Skip to content

fix: allow 30s for waits that depend on a backend round-trip - #53

Merged
saqibmanan merged 1 commit into
CIfrom
fix/backend-dependent-wait-timeouts
Sep 11, 2026
Merged

saqibmanan merged 1 commit into
CIfrom
fix/backend-dependent-wait-timeouts

Conversation

@saqibmanan

Copy link
Copy Markdown
Contributor

The post-create Metadata-tab wait (10s, 4 places) and select_combobox_option's option wait (default 15s) both wait on the backend — a create mutation plus navigation, or an option list populated by a GraphQL query. select_sdg_goals already allows 30s for the same kind of option list.

These timed out repeatedly in deploy-pipeline smoke runs while the target host was under load from three concurrent suites. Validated locally: full provider-smoke against dev, run concurrently with api and consumer suites, 12 passed, 1 xfailed, 1 xpassed, zero reruns — test_prv_002b/006/006b all passed first try.

The post-create Metadata-tab wait (10s) and select_combobox_option's
option wait (default 15s) both wait on the backend: a create mutation
plus navigation, or an option list populated by a GraphQL query on page
load. select_sdg_goals already allows 30s for exactly this kind of
option list. Both timed out repeatedly in deploy-pipeline smoke runs
while the target host was under concurrent load from three suites.
@saqibmanan
saqibmanan merged commit 8824aac into CI Sep 11, 2026
3 checks passed
@saqibmanan
saqibmanan deleted the fix/backend-dependent-wait-timeouts branch September 11, 2026 13:15
saqibmanan added a commit that referenced this pull request Sep 11, 2026
…ess (#56)

* fix: allow 30s for waits that depend on a backend round-trip (#53)

The post-create Metadata-tab wait (10s) and select_combobox_option's
option wait (default 15s) both wait on the backend: a create mutation
plus navigation, or an option list populated by a GraphQL query on page
load. select_sdg_goals already allows 30s for exactly this kind of
option list. Both timed out repeatedly in deploy-pipeline smoke runs
while the target host was under concurrent load from three suites.

* fix: skip dev-only GA-absent checks when the target isn't a dev host

HOME_URL_DEV is whichever site the suite targets. Pointed at prod, the
two dev GA tests asserted GA is absent on prod, where it is correctly
present (G-41CT4XZ341). They now skip unless the host starts with dev.,
matching the contract stated in the file header.

* fix: rerun on StaleElementReferenceException as a timing signature

A React re-render between wait_and_capture returning an element and the
caller's is_displayed() leaves a stale reference -- a race, not a logic
failure, and the same shape across ~43 wait_and_capture callers. Seen on
test_TC_SEC_02_search_bar against prod.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant