… after startup (#688)

* chore(inkless): align make targets with inkless ci

* chore(inkless): normalize ascii punctuation

* fix(inkless:consolidation): report cross-tier log start offset promptly after startup

Born-consolidated ListOffsets(EARLIEST) returns
COALESCE(remote_log_start_offset, log_start_offset). The leader publishes
remote_log_start_offset via CrossTierLogStartReporter, but the reporter's first
flush was gated by log.initial.task.delay.ms (default 30s). Until it fired,
EARLIEST returned the pruned log_start_offset instead of the true cross-tier
earliest, so a consumer with auto.offset.reset=earliest could skip records still
readable from remote for up to ~30s after every broker startup.

Schedule the reporter with its own report interval as the initial delay (like the
sibling inkless-file-cleaner / log-pruner tasks) instead of logInitialTaskDelayMs.

InklessConsolidatedDisklessTopicsTest now polls until EARLIEST converges before
the hard assertion, matching the reporter's async publish contract (the check
previously raced the first flush and failed on fast runners, e.g. main).

* chore(inkless:ci): report failed tests via junit.py and stop masking flaky PR failures

Inkless CI logged only a raw gradle exit code on failure. Parse the JUnit XML
with junit.py (as ci.yml/build.yml do) so the job summary lists failed tests, and
let its exit code drive the outcome. Set kafka.test.xml.output.dir for clean
report paths.

Set maxTestRetries=0 so a single failure fails a PR to main instead of being
retried away. Full build.yml parity on PRs to main stays intentionally out of
scope: PR CI runs focalized tests to stay fast.

… after startup (#688)

* chore(inkless): align make targets with inkless ci

* chore(inkless): normalize ascii punctuation

* fix(inkless:consolidation): report cross-tier log start offset promptly after startup

Born-consolidated ListOffsets(EARLIEST) returns
COALESCE(remote_log_start_offset, log_start_offset). The leader publishes
remote_log_start_offset via CrossTierLogStartReporter, but the reporter's first
flush was gated by log.initial.task.delay.ms (default 30s). Until it fired,
EARLIEST returned the pruned log_start_offset instead of the true cross-tier
earliest, so a consumer with auto.offset.reset=earliest could skip records still
readable from remote for up to ~30s after every broker startup.

Schedule the reporter with its own report interval as the initial delay (like the
sibling inkless-file-cleaner / log-pruner tasks) instead of logInitialTaskDelayMs.

InklessConsolidatedDisklessTopicsTest now polls until EARLIEST converges before
the hard assertion, matching the reporter's async publish contract (the check
previously raced the first flush and failed on fast runners, e.g. main).

* chore(inkless:ci): report failed tests via junit.py and stop masking flaky PR failures

Inkless CI logged only a raw gradle exit code on failure. Parse the JUnit XML
with junit.py (as ci.yml/build.yml do) so the job summary lists failed tests, and
let its exit code drive the outcome. Set kafka.test.xml.output.dir for clean
report paths.

Set maxTestRetries=0 so a single failure fails a PR to main instead of being
retried away. Full build.yml parity on PRs to main stays intentionally out of
scope: PR CI runs focalized tests to stay fast.

PR #688 removed all test retries from inkless.yml, exposing the CI to
known upstream flakiness (e.g. ReplicaManagerTest). Split the retry
policy by test ownership: kafka-owned suites are retried
(maxTestRetries=1, maxTestRetryFailures=3), while inkless-owned tests
(:storage:inkless and *Inkless*/*Diskless*/io.aiven.inkless.*) run
without retries so our own flakiness stays visible. The mixed :core:test
invocation is split accordingly.

Also add kafka.server.RequestQuotaTest to the kafka-owned :core:test
group. It iterates over every broker ApiKey, so it guards that each
inkless-added API (e.g. ALTER_DISKLESS_SWITCH) is wired into the
request-building/quota paths, shifting that check left from nightly to
per-PR CI.

…le (#692)

* fix(inkless:test): handle ALTER_DISKLESS_SWITCH in RequestQuotaTest

RequestQuotaTest iterates over all ApiKeys.brokerApis and builds a
request for each, but requestBuilder had no case for the inkless-added
ALTER_DISKLESS_SWITCH key, hitting the fallthrough and throwing
IllegalArgumentException. This deterministically failed
testUnauthorizedThrottle, testUnthrottledClient and
testResponseThrottleTime.

* fix(inkless:ci): correct nightly JUnit report path

The nightly parse step pointed junit.py at build/junit-xml/<java>, but
copyTestXml writes to build/junit-xml/<module>/<java>. The glob matched
nothing, so real test failures were reported as '0 tests' and the job
failed only on the gradle exit code. Point at build/junit-xml so the
recursive **/*.xml glob finds all module reports, matching inkless.yml.

* fix(inkless:ci): retry kafka-owned tests, keep inkless tests unretried

PR #688 removed all test retries from inkless.yml, exposing the CI to
known upstream flakiness (e.g. ReplicaManagerTest). Split the retry
policy by test ownership: kafka-owned suites are retried
(maxTestRetries=1, maxTestRetryFailures=3), while inkless-owned tests
(:storage:inkless and *Inkless*/*Diskless*/io.aiven.inkless.*) run
without retries so our own flakiness stays visible. The mixed :core:test
invocation is split accordingly.

Also add kafka.server.RequestQuotaTest to the kafka-owned :core:test
group. It iterates over every broker ApiKey, so it guards that each
inkless-added API (e.g. ALTER_DISKLESS_SWITCH) is wired into the
request-building/quota paths, shifting that check left from nightly to
per-PR CI.

* fix(inkless:ci): enforce the test timeout across the whole suite chain

The test step chained gradle invocations with && but only wrapped the
first call in 'timeout ${TIMEOUT_MINUTES}m', so a hang in any later
suite ran untimed and never produced exit 124 (which gates the
thread-dump upload). Wrap the entire chain in a single
'timeout ... bash -c' so the budget covers all suites and a hang
anywhere is killed and reported. Applied to inkless.yml and
inkless-nightly.yml.

…le (#692)

* fix(inkless:test): handle ALTER_DISKLESS_SWITCH in RequestQuotaTest

RequestQuotaTest iterates over all ApiKeys.brokerApis and builds a
request for each, but requestBuilder had no case for the inkless-added
ALTER_DISKLESS_SWITCH key, hitting the fallthrough and throwing
IllegalArgumentException. This deterministically failed
testUnauthorizedThrottle, testUnthrottledClient and
testResponseThrottleTime.

* fix(inkless:ci): correct nightly JUnit report path

The nightly parse step pointed junit.py at build/junit-xml/<java>, but
copyTestXml writes to build/junit-xml/<module>/<java>. The glob matched
nothing, so real test failures were reported as '0 tests' and the job
failed only on the gradle exit code. Point at build/junit-xml so the
recursive **/*.xml glob finds all module reports, matching inkless.yml.

* fix(inkless:ci): retry kafka-owned tests, keep inkless tests unretried

PR #688 removed all test retries from inkless.yml, exposing the CI to
known upstream flakiness (e.g. ReplicaManagerTest). Split the retry
policy by test ownership: kafka-owned suites are retried
(maxTestRetries=1, maxTestRetryFailures=3), while inkless-owned tests
(:storage:inkless and *Inkless*/*Diskless*/io.aiven.inkless.*) run
without retries so our own flakiness stays visible. The mixed :core:test
invocation is split accordingly.

Also add kafka.server.RequestQuotaTest to the kafka-owned :core:test
group. It iterates over every broker ApiKey, so it guards that each
inkless-added API (e.g. ALTER_DISKLESS_SWITCH) is wired into the
request-building/quota paths, shifting that check left from nightly to
per-PR CI.

* fix(inkless:ci): enforce the test timeout across the whole suite chain

The test step chained gradle invocations with && but only wrapped the
first call in 'timeout ${TIMEOUT_MINUTES}m', so a hang in any later
suite ran untimed and never produced exit 124 (which gates the
thread-dump upload). Wrap the entire chain in a single
'timeout ... bash -c' so the budget covers all suites and a hang
anywhere is killed and reported. Applied to inkless.yml and
inkless-nightly.yml.

…le (#692)

* fix(inkless:test): handle ALTER_DISKLESS_SWITCH in RequestQuotaTest

RequestQuotaTest iterates over all ApiKeys.brokerApis and builds a
request for each, but requestBuilder had no case for the inkless-added
ALTER_DISKLESS_SWITCH key, hitting the fallthrough and throwing
IllegalArgumentException. This deterministically failed
testUnauthorizedThrottle, testUnthrottledClient and
testResponseThrottleTime.

* fix(inkless:ci): correct nightly JUnit report path

The nightly parse step pointed junit.py at build/junit-xml/<java>, but
copyTestXml writes to build/junit-xml/<module>/<java>. The glob matched
nothing, so real test failures were reported as '0 tests' and the job
failed only on the gradle exit code. Point at build/junit-xml so the
recursive **/*.xml glob finds all module reports, matching inkless.yml.

* fix(inkless:ci): retry kafka-owned tests, keep inkless tests unretried

PR #688 removed all test retries from inkless.yml, exposing the CI to
known upstream flakiness (e.g. ReplicaManagerTest). Split the retry
policy by test ownership: kafka-owned suites are retried
(maxTestRetries=1, maxTestRetryFailures=3), while inkless-owned tests
(:storage:inkless and *Inkless*/*Diskless*/io.aiven.inkless.*) run
without retries so our own flakiness stays visible. The mixed :core:test
invocation is split accordingly.

Also add kafka.server.RequestQuotaTest to the kafka-owned :core:test
group. It iterates over every broker ApiKey, so it guards that each
inkless-added API (e.g. ALTER_DISKLESS_SWITCH) is wired into the
request-building/quota paths, shifting that check left from nightly to
per-PR CI.

* fix(inkless:ci): enforce the test timeout across the whole suite chain

The test step chained gradle invocations with && but only wrapped the
first call in 'timeout ${TIMEOUT_MINUTES}m', so a hang in any later
suite ran untimed and never produced exit 124 (which gates the
thread-dump upload). Wrap the entire chain in a single
'timeout ... bash -c' so the budget covers all suites and a hang
anywhere is killed and reported. Applied to inkless.yml and
inkless-nightly.yml.