fix: dedupe aliases before final request

[shruthilayaj]

The preflight was returning duplicated tag key names (we store the same
attribute key in different format in EAP). Dedupe it before making the
second request.

[cursor]

Bug: Pagination never detects next page after limit change

The pagination has_more check is broken. Previously data_fn was called with limit + 1 to detect if more results exist, but now it's called with just limit. However, the has_more condition still checks if data[1] >= limit + 1. Since data_fn returns len(request_attrs_list) which is at most limit (due to the slice [offset : offset + limit]), this condition will never be true. As a result, pagination will never show a "next page" link even when more results exist.

 

[cursor]

Bug: Pagination never detects next page after limit change

The pagination has_more check is broken. Previously data_fn was called with limit + 1 to detect if more results exist, but now it's called with just limit. However, the has_more condition still checks if data[1] >= limit + 1. Since data_fn returns len(request_attrs_list) which is at most limit (due to the slice [offset : offset + limit]), this condition will never be true. As a result, pagination will never show a "next page" link even when more results exist.

 

[codecov]

❌ 2 Tests Failed:

Tests completed	Failed	Passed	Skipped
30553	2	30551	242

View the top 2 failed test(s) by shortest run time

tests.snuba.api.endpoints.test_organization_trace_item_stats.OrganizationTraceItemsStatsEndpointTest::test_custom_limit_parameter

Stack Traces | 15.9s run time

#x1B[1m#x1B[.../api/endpoints/test_organization_trace_item_stats.py#x1B[0m:235: in test_custom_limit_parameter
    assert (
#x1B[1m#x1B[31mE   AssertionError: assert 1 == 2#x1B[0m
#x1B[1m#x1B[31mE    +  where 1 = len({'sentry.custom3': [{'label': 'value3', 'value': 1.0}]})#x1B[0m

tests.snuba.api.endpoints.test_organization_trace_item_stats.OrganizationTraceItemsStatsEndpointTest::test_substring_match_returns_known_public_aliases

Stack Traces | 18.5s run time

#x1B[1m#x1B[.../api/endpoints/test_organization_trace_item_stats.py#x1B[0m:145: in test_substring_match_returns_known_public_aliases
    assert len(response.data["data"]) == 1
#x1B[1m#x1B[31mE   assert 0 == 1#x1B[0m
#x1B[1m#x1B[31mE    +  where 0 = len([])#x1B[0m

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}

[getsantry]

This pull request has gone three weeks without activity. In another week, I will close it.

But! If you comment or otherwise update it, I will reset the clock, and if you add the label WIP, I will leave it alone unless WIP is removed ... forever!

---

"A weed is but an unloved flower." ― Ella Wheeler Wilcox 🥀

[cursor]

Deduplication tracks internal names instead of public aliases

Medium Severity

The deduplication logic adds internal_name to sanitized_keys_set and checks against it, but internal_alias_attr_keys comes from dictionary keys which are already unique. The PR's goal is to deduplicate when different internal names map to the same public_alias, but the set tracks internal_name instead of public_alias. This means multiple internal names mapping to the same public alias will all be included, defeating the deduplication purpose.