def test_invalid_add_array(self):
arr = [1, 2, 3]
inv1 = [1, [2, 3]]
inv2 = []
with pytest.raises(APIError):
op = InOperator('certname')
op.add_array(inv1)
with pytest.raises(APIError):
op = InOperator('certname')
op.add_array(inv2)
with pytest.raises(APIError):
op = InOperator('certname')
op.add_array(arr)
op.add_array(arr)
with pytest.raises(APIError):
op = InOperator('certname')
op.add_array(arr)
ex = ExtractOperator()
ex.add_field("certname")
op.add_query(ex)
def test_fromoperator(self):
op = InOperator('certname')
ex = ExtractOperator()
ex.add_field(["certname", "facts"])
fr = FromOperator("facts")
fr.add_query(ex)
fr.add_offset(10)
op.add_query(fr)
assert repr(op) == 'Query: ["in", "certname", ' \
'["from", "facts", ["extract", ' \
'["certname", "facts"]], ["offset", 10]]]'
# last example on page
# https://puppet.com/docs/puppetdb/5.1/api/query/v4/ast.html
op = InOperator('certname')
ex = ExtractOperator()
ex.add_field('certname')
fr = FromOperator('fact_contents')
nd = AndOperator()
nd.add(
EqualsOperator("path", ["networking", "eth0", "macaddresses", 0]))
nd.add(EqualsOperator("value", "aa:bb:cc:dd:ee:00"))
ex.add_query(nd)
fr.add_query(ex)
op.add_query(fr)
assert str(op) == '["in", "certname", ' \
'["from", "fact_contents", ' \
'["extract", ["certname"], ["and", ["=", "path", ' \
'["networking", "eth0", "macaddresses", 0]], ' \
'["=", "value", "aa:bb:cc:dd:ee:00"]]]]]'
def test_with_add_function_operator(self):
op = ExtractOperator()
op.add_field(FunctionOperator('to_string',
'producer_timestamp',
'FMDAY'))
op.add_field(FunctionOperator('count'))
op.add_group_by(FunctionOperator('to_string',
'producer_timestamp',
'FMDAY'))
assert str(op) == '["extract", '\
'[["function", "to_string", "producer_timestamp", "FMDAY"], '\
'["function", "count"]], '\
'["group_by", '\
'["function", "to_string", "producer_timestamp", "FMDAY"]]]'
assert repr(op) == 'Query: ["extract", '\
'[["function", "to_string", "producer_timestamp", "FMDAY"], '\
'["function", "count"]], '\
'["group_by", '\
'["function", "to_string", "producer_timestamp", "FMDAY"]]]'
assert str(op) == '["extract", ' \
'[["function", "to_string", "producer_timestamp", "FMDAY"], '\
'["function", "count"]], '\
'["group_by", '\
'["function", "to_string", "producer_timestamp", "FMDAY"]]]'
def test_events_endpoint(self):
assert str(InOperator('certname')) == \
'["in", "certname"]'
op = InOperator('certname')
ex = ExtractOperator()
ex.add_field("certname")
op.add_query(ex)
assert repr(op) == 'Query: ["in", "certname", ' \
'["extract", ["certname"]]]'
def main():
"""main entry point"""
args = get_args()
logging.basicConfig(level=get_log_level(args.verbose))
failed_nodes = []
pdb = connect()
nodes = defaultdict(dict)
max_age = datetime.utcnow() - timedelta(hours=args.max_age)
extract = ExtractOperator()
extract.add_field(['certname', FunctionOperator('count'), 'status'])
extract.add_group_by(['certname', 'status'])
extract.add_query(GreaterOperator('receive_time', max_age.isoformat()))
# pypuppetdb does have a `reports` method which wraps `_query`. however it yields
# results of type pypuppetdb.Report which expects a number of parameters e.g. hash
# to be present in the result payload. however we don't extract theses values and
# therefore have to resort to the more powerful private method
reports = pdb._query('reports', query=extract) # pylint: disable=protected-access
for report in reports:
nodes[report['certname']][report['status']] = report['count']
if args.dev:
failed_nodes = [hostname for hostname, node in nodes.items()
if not node.get('unchanged', 0)]
else:
for fqdn, node in nodes.items():
# skip hosts with no unchanged reports:
if node.get('unchanged', 0):
continue
# skip staging servers:
# - hostname starting labstest*
# - hostname ending dev or dev\d{4}
# - hostname ending test or test\d{4}
if (fqdn.startswith('labtest')
or search(r'(:?dev|test)(:?\d{4})?$', fqdn.split('.')[0]) is not None):
logger.debug('%s: Skipping staging host', fqdn)
continue
failed_nodes.append(fqdn)
if len(failed_nodes) >= args.critical:
print('CRITICAL: the following ({}) node(s) change every puppet run: {}'.format(
len(failed_nodes), ', '.join(failed_nodes)))
return 2
if len(failed_nodes) >= args.warning:
print('WARNING: the following ({}) node(s) change every puppet run: {}'.format(
len(failed_nodes), ', '.join(failed_nodes)))
return 1
print('OK: all nodes running as expected')
return 0
def _build_query(env, start, end, certname=None):
"""Build a extract query with optional certname and environment."""
query = ExtractOperator()
query.add_field(FunctionOperator('count'))
query.add_field('status')
subquery = AndOperator()
subquery.add(GreaterEqualOperator('producer_timestamp', start))
subquery.add(LessOperator('producer_timestamp', end))
if certname is not None:
subquery.add(EqualsOperator('certname', certname))
if env != '*':
subquery.add(EqualsOperator('environment', env))
query.add_query(subquery)
query.add_group_by("status")
return query
def _build_query(env, start, end, certname=None):
"""Build a extract query with optional certname and environment."""
query = ExtractOperator()
query.add_field(FunctionOperator('count'))
query.add_field('status')
subquery = AndOperator()
subquery.add(GreaterEqualOperator('start_time', start))
subquery.add(LessOperator('start_time', end))
if certname is not None:
subquery.add(EqualsOperator('certname', certname))
if env != '*':
subquery.add(EqualsOperator('environment', env))
query.add_query(subquery)
query.add_group_by("status")
return query
def main():
"""main entry point"""
args = get_args()
logging.basicConfig(level=get_log_level(args.verbose))
pdb = connect()
nodes = defaultdict(dict)
max_age = datetime.utcnow() - timedelta(hours=args.max_age)
extract = ExtractOperator()
extract.add_field(['certname', FunctionOperator('count'), 'status'])
extract.add_group_by(['certname', 'status'])
extract.add_query(GreaterOperator('receive_time', max_age.isoformat()))
# pypuppetdb does have a `reports` method which wraps `_query`. however it yields
# results of type pypuppetdb.Report which expects a number of parameters e.g. hash
# to be present in the result payload. however we don't extract theses values and
# therefore have to resort to the more powerful private method
reports = pdb._query('reports', query=extract) # pylint: disable=protected-access
for report in reports:
nodes[report['certname']][report['status']] = report['count']
failed_nodes = [
hostname for hostname, node in nodes.items()
if not node.get('unchanged', 0)
]
if len(failed_nodes) >= args.critical:
print(
'CRITICAL: the following ({}) node(s) change every puppet run: {}'.
format(len(failed_nodes), ', '.join(failed_nodes)))
return 2
if len(failed_nodes) >= args.warning:
print(
'WARNING: the following ({}) node(s) change every puppet run: {}'.
format(len(failed_nodes), ', '.join(failed_nodes)))
return 1
print('OK: all nodes running as expected')
return 0
def test_with_add_group_by(self):
op = ExtractOperator()
op.add_field(['certname', 'fact_environment', 'catalog_environment'])
op.add_query(EqualsOperator('domain', 'example.com'))
op.add_group_by(["fact_environment", "catalog_environment"])
with pytest.raises(APIError):
op.add_group_by({"deactivated": False})
assert repr(op) == 'Query: ["extract", '\
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"], '\
'["group_by", "fact_environment", "catalog_environment"]]'
assert str(op) == '["extract", '\
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"], '\
'["group_by", "fact_environment", "catalog_environment"]]'
assert str(op) == '["extract", ' \
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"], '\
'["group_by", "fact_environment", "catalog_environment"]]'
def test_with_add_query(self):
op = ExtractOperator()
op.add_field(['certname', 'fact_environment', 'catalog_environment'])
with pytest.raises(APIError):
op.add_query({'less': 42, 'greater': 50})
op.add_query(EqualsOperator('domain', 'example.com'))
assert repr(op) == 'Query: ["extract", '\
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"]]'
assert str(op) == '["extract", '\
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"]]'
assert str(op) == '["extract", ' \
'["certname", "fact_environment", "catalog_environment"], '\
'["=", "domain", "example.com"]]'
with pytest.raises(APIError):
op.add_query(GreaterOperator("processorcount", 1))
def test_add_query(self):
fr = FromOperator("facts")
op = EqualsOperator("certname", "test01")
fr.add_query(op)
assert str(fr) == '["from", "facts", ["=", "certname", "test01"]]'
fr2 = FromOperator("facts")
op2 = "test, test, test"
with pytest.raises(APIError):
fr2.add_query(op2)
fr2.add_query(op)
with pytest.raises(APIError):
fr2.add_query(op)
fr3 = FromOperator("facts")
op3 = ExtractOperator()
op3.add_field(['certname', 'fact_environment', 'catalog_environment'])
fr3.add_query(op3)
assert str(fr3) == \
'["from", "facts", ["extract", '\
'["certname", "fact_environment", "catalog_environment"]]]'
def get_puppetdb_resources():
"""Get a list of unique resource inuse by the puppetdb
This functions assumes one can connecto puppetdb:8080 i.e.
add the following to /etc/hosts
127.0.0.1 puppetdb
and:
ssh -L8080:localhost:8080 puppetdb1002.eqiad.wmnet
"""
unique_resources = set()
db = connect()
extract = ExtractOperator()
extract.add_field(['type', 'title'])
# TODO: this is pretty slow as it gets all resources then dose a unique
# would be better to do a select distinct via the api
resources = db._query('resources', query=extract)
for resource in resources:
# if we have a class we want the title to know which class
if resource['type'] == 'Class':
unique_resources.add(resource['title'].lower())
else:
unique_resources.add(resource['type'].lower())
return unique_resources
def test_with_add_field(self):
op = ExtractOperator()
with pytest.raises(APIError):
repr(op)
with pytest.raises(APIError):
str(op)
with pytest.raises(APIError):
str(op)
op.add_field("certname")
op.add_field(['fact_environment', 'catalog_environment'])
assert repr(op) == 'Query: ["extract", '\
'["certname", "fact_environment", "catalog_environment"]]'
assert str(op) == '["extract", '\
'["certname", "fact_environment", "catalog_environment"]]'
assert str(op) == '["extract", ' \
'["certname", "fact_environment", "catalog_environment"]]'
with pytest.raises(APIError):
op.add_field({'equal': 'operatingsystemrelease'})
def index(env):
"""This view generates the index page and displays a set of metrics and
latest reports on nodes fetched from PuppetDB.
:param env: Search for nodes in this (Catalog and Fact) environment
:type env: :obj:`string`
"""
envs = environments()
metrics = {'num_nodes': 0, 'num_resources': 0, 'avg_resources_node': 0}
check_env(env, envs)
if env == '*':
query = app.config['OVERVIEW_FILTER']
prefix = 'puppetlabs.puppetdb.population'
db_version = get_db_version(puppetdb)
query_type, metric_version = metric_params(db_version)
num_nodes = get_or_abort(puppetdb.metric,
"{0}{1}".format(
prefix, ':%sname=num-nodes' % query_type),
version=metric_version)
num_resources = get_or_abort(puppetdb.metric,
"{0}{1}".format(
prefix,
':%sname=num-resources' % query_type),
version=metric_version)
metrics['num_nodes'] = num_nodes['Value']
metrics['num_resources'] = num_resources['Value']
try:
# Compute our own average because avg_resources_node['Value']
# returns a string of the format "num_resources/num_nodes"
# example: "1234/9" instead of doing the division itself.
metrics['avg_resources_node'] = "{0:10.0f}".format(
(num_resources['Value'] / num_nodes['Value']))
except ZeroDivisionError:
metrics['avg_resources_node'] = 0
else:
query = AndOperator()
query.add(EqualsOperator('catalog_environment', env))
num_nodes_query = ExtractOperator()
num_nodes_query.add_field(FunctionOperator('count'))
num_nodes_query.add_query(query)
if app.config['OVERVIEW_FILTER'] is not None:
query.add(app.config['OVERVIEW_FILTER'])
num_resources_query = ExtractOperator()
num_resources_query.add_field(FunctionOperator('count'))
num_resources_query.add_query(EqualsOperator("environment", env))
num_nodes = get_or_abort(puppetdb._query,
'nodes',
query=num_nodes_query)
num_resources = get_or_abort(puppetdb._query,
'resources',
query=num_resources_query)
metrics['num_nodes'] = num_nodes[0]['count']
metrics['num_resources'] = num_resources[0]['count']
try:
metrics['avg_resources_node'] = "{0:10.0f}".format(
(num_resources[0]['count'] / num_nodes[0]['count']))
except ZeroDivisionError:
metrics['avg_resources_node'] = 0
nodes = get_or_abort(puppetdb.nodes,
query=query,
unreported=app.config['UNRESPONSIVE_HOURS'],
with_status=True,
with_event_numbers=app.config['WITH_EVENT_NUMBERS'])
nodes_overview = []
stats = {
'changed': 0,
'unchanged': 0,
'failed': 0,
'unreported': 0,
'noop': 0
}
for node in nodes:
if node.status == 'unreported':
stats['unreported'] += 1
elif node.status == 'changed':
stats['changed'] += 1
elif node.status == 'failed':
stats['failed'] += 1
elif node.status == 'noop':
stats['noop'] += 1
else:
stats['unchanged'] += 1
if node.status != 'unchanged':
nodes_overview.append(node)
return render_template('index.html',
metrics=metrics,
nodes=nodes_overview,
stats=stats,
envs=envs,
current_env=env)
def radiator(env):
"""This view generates a simplified monitoring page
akin to the radiator view in puppet dashboard
"""
envs = environments()
check_env(env, envs)
if env == '*':
db_version = get_db_version(puppetdb)
query_type, metric_version = metric_params(db_version)
query = None
metrics = get_or_abort(
puppetdb.metric,
'puppetlabs.puppetdb.population:%sname=num-nodes' % query_type,
version=metric_version)
num_nodes = metrics['Value']
else:
query = AndOperator()
metric_query = ExtractOperator()
query.add(EqualsOperator("catalog_environment", env))
metric_query.add_field(FunctionOperator('count'))
metric_query.add_query(query)
metrics = get_or_abort(puppetdb._query, 'nodes', query=metric_query)
num_nodes = metrics[0]['count']
nodes = puppetdb.nodes(query=query,
unreported=app.config['UNRESPONSIVE_HOURS'],
with_status=True)
stats = {
'changed_percent': 0,
'changed': 0,
'failed_percent': 0,
'failed': 0,
'noop_percent': 0,
'noop': 0,
'skipped_percent': 0,
'skipped': 0,
'unchanged_percent': 0,
'unchanged': 0,
'unreported_percent': 0,
'unreported': 0,
}
for node in nodes:
if node.status == 'unreported':
stats['unreported'] += 1
elif node.status == 'changed':
stats['changed'] += 1
elif node.status == 'failed':
stats['failed'] += 1
elif node.status == 'noop':
stats['noop'] += 1
elif node.status == 'skipped':
stats['skipped'] += 1
else:
stats['unchanged'] += 1
try:
stats['changed_percent'] = int(100 *
(stats['changed'] / float(num_nodes)))
stats['failed_percent'] = int(100 * stats['failed'] / float(num_nodes))
stats['noop_percent'] = int(100 * stats['noop'] / float(num_nodes))
stats['skipped_percent'] = int(100 *
(stats['skipped'] / float(num_nodes)))
stats['unchanged_percent'] = int(
100 * (stats['unchanged'] / float(num_nodes)))
stats['unreported_percent'] = int(
100 * (stats['unreported'] / float(num_nodes)))
except ZeroDivisionError:
stats['changed_percent'] = 0
stats['failed_percent'] = 0
stats['noop_percent'] = 0
stats['skipped_percent'] = 0
stats['unchanged_percent'] = 0
stats['unreported_percent'] = 0
if ('Accept' in request.headers
and request.headers["Accept"] == 'application/json'):
return jsonify(**stats)
return render_template('radiator.html', stats=stats, total=num_nodes)