ci-timing: weekly CI timing report

[github-actions[bot]]

Reports from the job this thread is named after, one comment per report. Posted by util/discussion_report.py (#2030).

[github-actions[bot]]

CI timing, 2026-10-19T07:49Z to 2026-10-26T07:49Z

Jobs read for an evenly spaced sample of each workflow's completed runs: unit-tests.yml 301 of 559 (2026-10-19T08:01Z to 2026-10-26T06:32Z); python-compatibility.yml 60 of 565 (2026-10-19T08:01Z to 2026-10-26T06:56Z); lint.yml 60 of 356 (2026-10-19T08:01Z to 2026-10-26T06:56Z); codeql-analysis.yml 60 of 565 (2026-10-19T08:01Z to 2026-10-26T06:56Z); deploy-pages.yml 60 of 556 (2026-10-19T08:01Z to 2026-10-26T07:04Z); experimental.yml 60 of 560 (2026-10-19T08:01Z to 2026-10-26T06:45Z).

• HARD H2: Plain unittest ordering (project) step Run tests/project and the root-level modules under plain unittest p90 is 11.8 min, at or above 50% of its 20-minute step cap.
• HARD H4: Plain unittest ordering (project) step Run tests/project and the root-level modules under plain unittest reaches its 20-minute step cap in 7.3 days on a linear fit.
• WARN H1: Python graalpy-25.4 (experimental) (experimental.yml) p90 is 76.2 min, at or above 50% of its 140-minute cap. Non-blocking, so a warning (ci(python): the GraalPy experimental leg reaches 26% of the suite before its 25-minute cap #1732).
• WARN H2: Python graalpy-25.4 (experimental) step Run unit tests p90 is 73.0 min, at or above 50% of its 120-minute step cap. Non-blocking, so a warning (ci(python): the GraalPy experimental leg reaches 26% of the suite before its 25-minute cap #1732).
• WARN W9: pull-request time to Required checks passed p90 is 152.6 min, at or above 30; HARD (H3) only as a rise on the baseline.
• WARN W2: pull-request time to mergeable median is 52.8 min, above 18.
• WARN W3: Python 3.15's Run unit tests takes 3.16x its paired legs' (median), above 1.8.
• WARN W5: pull_request run queue p90 is 34.2 min, above 5.
• WARN W5: push run queue p90 is 81.2 min, above 15.
• WARN W7: 1 module(s) over 30 s summed on 3.13: tests.test_tier_guard
• WARN W7: 7 test(s) over 10 s on 3.13: tests.test_tier_guard.WorkflowAgreementTests::test_a_leg_that_asks_for_the_wrong_list_is_caught, tests.test_tier_guard.WorkflowAgreementTests::test_the_legs_leave_out_exactly_the_latest_tier_or_the_pins, tests.test_tier_guard.WorkflowAgreementTests::test_a_leg_that_stops_ignoring_the_latest_tier_is_caught
• WARN W7: 12 class(es) whose every test costs about as much as the first, the per-test re-import signature: tests.const.test_const_apptype_lazy_1538_unit.LazyAppTypeContractTests, tests.const.test_const_apptype_lazy_1538_unit.LazyAppTypeImportTests, tests.const.test_const_enum_builtin_parity.ConstEnumRegisterFallbackTests, tests.corekit.test_enum_lookup_reparent_877_unit.NoMintingTests, tests.foundation.test_extraction_no_eof.NoEOFTerminatesOnAnExhaustedInput

Unit Tests runs

Event	Listed	Cancelled	Measured	Queue p50 / p90	To mergeable p50 / p90	Wall p50 / p90	Runner-min per run	Jobs per run
pull_request	350	108	109	3.0 / 34.2	52.8 / 152.6	52.8 / 161.0	120.5	32
push	209	1	151	6.7 / 81.2	101.2 / 265.3	110.9 / 274.7	172.3	43

Superseded (cancelled) pull-request runs: 108 of 350 (31%), about 3483 runner-min, scaled from the 41 read.

Jobs, by p90 share of cap (top 15)

Workflow	Job	n	Median	p90	Max	Queue p90	Cap	p90 / cap	Trend min/day	Days to 50%	Days to cap
experimental.yml	Python graalpy-25.4 (experimental)	44	45.8	76.2	113.5	28.0	140	54%	-4.16	-	-
unit-tests.yml	Python 3.15	261	19.6	21.7	26.0	79.3	45	48%	-0.10	-	-
unit-tests.yml	Python pypy3.10 (KTLO)	181	11.3	14.3	15.4	73.1	45	32%	2.16	3.7	14.1
unit-tests.yml	Python pypy3.11	175	12.7	13.9	15.0	79.6	45	31%	0.26	39.0	126.0
unit-tests.yml	Plain unittest ordering (project)	151	9.9	12.6	13.8	103.0	45	28%	1.08	8.9	29.9
unit-tests.yml	Python 3.12	266	6.3	11.8	16.0	74.3	45	26%	-0.67	-	-
unit-tests.yml	Python 3.14	266	6.3	11.1	14.4	77.4	45	25%	-0.62	-	-
unit-tests.yml	Python 3.11	268	5.9	10.8	15.8	73.8	45	24%	-0.66	-	-
unit-tests.yml	Python 3.13	264	6.1	10.8	15.4	79.6	45	24%	-0.66	-	-
unit-tests.yml	Project tests	245	3.5	4.7	14.0	63.8	20	24%	0.45	11.2	33.3
unit-tests.yml	Python 3.10 (KTLO)	267	6.1	10.3	14.9	74.4	45	23%	-0.50	-	-
unit-tests.yml	Python 3.9 (KTLO)	271	6.2	10.1	13.4	53.2	45	23%	-0.51	-	-
unit-tests.yml	Python 3.8 (KTLO)	272	5.4	7.9	9.4	55.2	45	17%	-0.28	-	-
unit-tests.yml	Plain unittest ordering (protocols (rest))	151	7.1	7.8	9.4	104.0	45	17%	0.16	92.9	231.7
unit-tests.yml	Engines (PyShark)	267	3.9	4.3	5.6	74.0	30	14%	0.01	1046.5	2450.3

Steps, by p90 (top 15)

Job	Step	n	Median	p90	Step cap	Days to cap
Python graalpy-25.4 (experimental)	Run unit tests	44	42.9	73.0	120	-
Python 3.15	Run unit tests	261	17.9	20.1	-	-
Python pypy3.10 (KTLO)	Run unit tests	146	12.1	13.3	-	-
Python pypy3.11	Run unit tests	174	11.6	12.8	-	-
Project tests	Run tests/project and the root-level modules under plain unittest	2	12.4	12.7	15	-
Plain unittest ordering (project)	Run tests/project and the root-level modules under plain unittest	151	9.2	11.8	20	7.3
Python 3.12	Run unit tests	266	5.6	11.3	-	-
Python 3.14	Run unit tests	266	5.5	10.3	-	-
Python 3.11	Run unit tests	268	5.2	10.3	-	-
Python 3.13	Run unit tests	264	5.4	10.1	-	-
Python 3.10 (KTLO)	Run unit tests	267	5.4	9.7	-	-
Python 3.9 (KTLO)	Run unit tests	271	5.4	9.4	-	-
Python 3.8 (KTLO)	Run unit tests	272	4.6	7.2	-	-
Plain unittest ordering (protocols (rest))	Run tests/protocols and the root-level modules under plain unittest	151	6.2	6.9	20	80.1
Python 3.14t (experimental)	Run unit tests	44	3.9	5.7	120	-

Critical path to Required checks passed

Event	Runs	Last to finish	Share	Median slack to runner-up
pull_request	109	Python 3.15	68%	6.0 min
push	151	Python 3.15	92%	6.0 min

Derived

• Python 3.15 Run unit tests over its paired legs: median 3.16x, range 1.27-4.45, n=258.
• ms per test on 3.13: median 132.6, n=5.
• Tests on 3.13 (run 40262953152): 2308 testcases; the suite's tests= reads 14578, which counts subtests.

Tests on 3.13, median of 5 push run(s)

Module	Seconds
tests.test_tier_guard	82.0
tests.protocols.test_frame_layer_limit_1610_unit	23.8
tests.protocols.internet.test_ext_chain_limit_rebuild_1619_unit	19.4
tests.protocols.misc.test_pcapng_unit	18.7
tests.test_support_helpers	17.6
tests.utilities.test_exceptions_excepthook	11.6
tests.const.test_const_enum_builtin_parity	10.9
tests.foundation.test_extraction_no_eof	10.7
tests.toolkit.test_pyshark_unit	10.3
tests.const.test_const_apptype_lazy_1538_unit	10.1

Class	Seconds
tests.test_tier_guard.WorkflowAgreementTests	62.1
tests.protocols.test_frame_layer_limit_1610_unit.FrameLayerLimitTests	22.2
tests.protocols.internet.test_ext_chain_limit_rebuild_1619_unit.ExtensionHeaderChainRebuildTests	19.4
tests.test_support_helpers.TimeLimitTests	16.8
tests.protocols.misc.test_pcapng_unit.PCAPNGTruncatedFileTests	12.0
tests.protocols.test_option_generator_deadline_scale_1428_unit.GeneratorDeadlineScaleTests	8.1
tests.test_tier_guard.DependencyGateCoverageTests	7.6
tests.corekit.test_info_child_finalisation_1490_unit.ChildFinalisationTests	7.6
tests.const.test_const_enum_builtin_parity.ConstEnumRegisterFallbackTests	7.3
tests.corekit.test_ipv4interface_mask_unit.TestIPv4InterfaceMask	6.8

Tests on 3.14, median of 5 push run(s)

Module	Seconds
tests.test_tier_guard	101.9
tests.protocols.test_frame_layer_limit_1610_unit	32.5
tests.test_tier_guard_xdist	31.0
tests.foundation.test_extraction_non_seekable_stream_1586_unit	30.3
tests.protocols.internet.test_ext_chain_limit_rebuild_1619_unit	24.6
tests.protocols.misc.test_pcapng_unit	20.6
tests.test_support_helpers	18.2
tests.utilities.test_exceptions_excepthook	14.1
tests.toolkit.test_pyshark_unit	13.0
tests.const.test_const_apptype_lazy_1538_unit	12.3

Class	Seconds
tests.test_tier_guard.WorkflowAgreementTests	64.9
tests.foundation.test_extraction_non_seekable_stream_1586_unit.TestNonSeekableStream	30.3
tests.protocols.test_frame_layer_limit_1610_unit.FrameLayerLimitTests	30.1
tests.protocols.internet.test_ext_chain_limit_rebuild_1619_unit.ExtensionHeaderChainRebuildTests	24.6
tests.test_tier_guard_xdist.XdistSubprocessTests	22.7
tests.test_support_helpers.TimeLimitTests	16.8
tests.test_tier_guard.DependencyGateFalsifiabilityTests	14.4
tests.protocols.misc.test_pcapng_unit.PCAPNGTruncatedFileTests	13.5
tests.corekit.test_info_child_finalisation_1490_unit.ChildFinalisationTests	9.2
tests.corekit.test_ipv4interface_mask_unit.TestIPv4InterfaceMask	8.2

Matched no job in the workflow file, so judged against the 360-minute default cap: deploy-pages.yml: Docs gate / Verdict covers this commit, deploy-pages.yml: Verdict covers this commit
No baseline at prev/report.json, so H3, W1, W4, W7's new entrants and W8 were not evaluated.

API calls: 683 (rate limit remaining: 4275).

---

Posted from CI Timing run at commit 91147a182b63a1d70fa92c3cd5b916ca84b67443, which ended failure.