def test_osd_status_metrics(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
expected_metrics = [
'ceph.read_op_per_sec', 'ceph.write_op_per_sec', 'ceph.op_per_sec'
]
for osd, pct_used in [('osd1', 94), ('osd2', 95)]:
expected_tags = EXPECTED_TAGS + ['ceph_osd:%s' % osd]
aggregator.assert_metric('ceph.osd.pct_used',
value=pct_used,
count=1,
tags=expected_tags)
aggregator.assert_metric('ceph.num_full_osds',
value=1,
count=1,
tags=EXPECTED_TAGS)
aggregator.assert_metric('ceph.num_near_full_osds',
value=1,
count=1,
tags=EXPECTED_TAGS)
for pool in ['rbd', 'scbench']:
expected_tags = EXPECTED_TAGS + ['ceph_pool:%s' % pool]
for metric in expected_metrics:
aggregator.assert_metric(metric, count=1, tags=expected_tags)
def test_luminous_osd_full_metrics(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
aggregator.assert_metric('ceph.num_full_osds', value=1)
aggregator.assert_metric('ceph.num_near_full_osds', value=1)
def test_warn_health(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
for metric in EXPECTED_METRICS:
aggregator.assert_metric(metric, count=1, tags=EXPECTED_TAGS)
aggregator.assert_service_check('ceph.overall_status', status=Ceph.WARNING, tags=EXPECTED_SERVICE_TAGS)
def test_luminous_warn_health(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
config = copy.deepcopy(BASIC_CONFIG)
config["collect_service_check_for"] = ['OSD_NEARFULL', 'OSD_FULL']
ceph_check.check(config)
aggregator.assert_service_check('ceph.overall_status', status=Ceph.CRITICAL, tags=EXPECTED_SERVICE_TAGS)
aggregator.assert_service_check('ceph.osd_nearfull', status=Ceph.WARNING, tags=EXPECTED_SERVICE_TAGS)
aggregator.assert_service_check('ceph.osd_full', status=Ceph.CRITICAL, tags=EXPECTED_SERVICE_TAGS)
def test_luminous_ok_health(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
config = copy.deepcopy(BASIC_CONFIG)
config["collect_service_check_for"] = ['OSD_NEARFULL']
ceph_check.check(config)
aggregator.assert_service_check('ceph.overall_status', status=Ceph.OK)
aggregator.assert_service_check('ceph.osd_nearfull', status=Ceph.OK)
aggregator.assert_service_check('ceph.pool_app_not_enabled', count=0)
def test_osd_perf_with_osdstats(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
for osd in ['osd0', 'osd1', 'osd2']:
expected_tags = EXPECTED_TAGS + ['ceph_osd:%s' % osd]
for metric in ['ceph.commit_latency_ms', 'ceph.apply_latency_ms']:
aggregator.assert_metric(metric, count=1, tags=expected_tags)
def test_check(aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
for metric in EXPECTED_METRICS:
aggregator.assert_metric(metric, at_least=1)
for sc in EXPECTED_SERVICE_CHECKS:
aggregator.assert_service_check(sc, status=Ceph.OK, tags=EXPECTED_SERVICE_TAGS)
aggregator.assert_service_check('ceph.overall_status', status=Ceph.OK, tags=EXPECTED_SERVICE_TAGS)
def test_osd_status_metrics_non_osd_health(_, aggregator):
"""
The `detail` key of `health detail` can contain info on the health of non-osd units:
shouldn't make the check fail
"""
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
aggregator.assert_metric('ceph.num_full_osds', value=0, count=1, tags=EXPECTED_TAGS)
aggregator.assert_metric('ceph.num_near_full_osds', value=0, count=1, tags=EXPECTED_TAGS)
def test_tagged_metrics(_, aggregator):
ceph_check = Ceph(CHECK_NAME, {}, {})
ceph_check.check(copy.deepcopy(BASIC_CONFIG))
for osd in ['osd0', 'osd1', 'osd2']:
expected_tags = EXPECTED_TAGS + ['ceph_osd:%s' % osd]
for metric in ['ceph.commit_latency_ms', 'ceph.apply_latency_ms']:
aggregator.assert_metric(metric, count=1, tags=expected_tags)
for pool in ['pool0', 'rbd']:
expected_tags = EXPECTED_TAGS + ['ceph_pool:%s' % pool]
for metric in EXPECTED_METRICS_POOL_TAGS:
aggregator.assert_metric(metric, count=1, tags=expected_tags)