def check_user_image_inline(user_id, image_id, tag, bundle):
"""
Execute a policy evaluation using the info in the request body including the bundle content
:param user_id:
:param image_id:
:param tag:
:param bundle:
:return:
"""
timer = time.time()
db = get_session()
cache_mgr = None
try:
# Input validation
if tag is None:
# set tag value to a value that only matches wildcards
tag = "*/*:*"
try:
img_obj = db.query(Image).get((image_id, user_id))
except:
return make_response_error("Image not found", in_httpcode=404), 404
if not img_obj:
log.info(
"Request for evaluation of image that cannot be found: user_id = {}, image_id = {}"
.format(user_id, image_id))
return make_response_error("Image not found", in_httpcode=404), 404
if evaluation_cache_enabled:
timer2 = time.time()
try:
try:
conn_timeout = (
ApiRequestContextProxy.get_service().configuration.get(
"catalog_client_conn_timeout",
DEFAULT_CACHE_CONN_TIMEOUT))
read_timeout = (
ApiRequestContextProxy.get_service().configuration.get(
"catalog_client_read_timeout",
DEFAULT_CACHE_READ_TIMEOUT))
cache_mgr = EvaluationCacheManager(img_obj, tag, bundle,
conn_timeout,
read_timeout)
except ValueError as err:
log.warn(
"Could not leverage cache due to error in bundle data: {}"
.format(err))
cache_mgr = None
if cache_mgr is None:
log.info(
"Could not initialize cache manager for policy evaluation, skipping cache usage"
)
else:
cached_result = cache_mgr.refresh()
if cached_result:
metrics.counter_inc(
name="anchore_policy_evaluation_cache_hits")
metrics.histogram_observe(
"anchore_policy_evaluation_cache_access_latency",
time.time() - timer2,
status="hit",
)
log.info(
"Returning cached result of policy evaluation for {}/{}, with tag {} and bundle {} with digest {}. Last evaluation: {}"
.format(
user_id,
image_id,
tag,
cache_mgr.bundle_id,
cache_mgr.bundle_digest,
cached_result.get("last_modified"),
))
return cached_result
else:
metrics.counter_inc(
name="anchore_policy_evaluation_cache_misses")
metrics.histogram_observe(
"anchore_policy_evaluation_cache_access_latency",
time.time() - timer2,
status="miss",
)
log.info(
"Policy evaluation not cached, or invalid, executing evaluation for {}/{} with tag {} and bundle {} with digest {}"
.format(
user_id,
image_id,
tag,
cache_mgr.bundle_id,
cache_mgr.bundle_digest,
))
except Exception as ex:
log.exception(
"Unexpected error operating on policy evaluation cache. Skipping use of cache."
)
else:
log.info("Policy evaluation cache disabled. Executing evaluation")
# Build bundle exec.
problems = []
executable_bundle = None
try:
# Allow deprecated gates here to support upgrade cases from old policy bundles.
executable_bundle = build_bundle(bundle,
for_tag=tag,
allow_deprecated=True)
if executable_bundle.init_errors:
problems = executable_bundle.init_errors
except InitializationError as e:
log.exception(
"Bundle construction and initialization returned errors")
problems = e.causes
eval_result = None
if not problems:
# Execute bundle
try:
eval_result = executable_bundle.execute(
img_obj, tag,
ExecutionContext(db_session=db, configuration={}))
except Exception as e:
log.exception(
"Error executing policy bundle {} against image {} w/tag {}: {}"
.format(bundle["id"], image_id, tag, e))
return (
make_response_error(
"Internal bundle evaluation error",
details={
"message":
"Cannot execute given policy against the image due to errors executing the policy bundle: {}"
.format(e)
},
in_httpcode=500,
),
500,
)
else:
# Construct a failure eval with details on the errors and mappings to send to client
eval_result = build_empty_error_execution(img_obj,
tag,
executable_bundle,
errors=problems,
warnings=[])
if (executable_bundle and executable_bundle.mapping
and len(executable_bundle.mapping.mapping_rules) == 1):
eval_result.executed_mapping = executable_bundle.mapping.mapping_rules[
0]
resp = PolicyEvaluation()
resp.user_id = user_id
resp.image_id = image_id
resp.tag = tag
resp.bundle = bundle
resp.matched_mapping_rule = (eval_result.executed_mapping.json() if
eval_result.executed_mapping else False)
resp.last_modified = int(time.time())
resp.final_action = eval_result.bundle_decision.final_decision.name
resp.final_action_reason = eval_result.bundle_decision.reason
resp.matched_whitelisted_images_rule = (
eval_result.bundle_decision.whitelisted_image.json()
if eval_result.bundle_decision.whitelisted_image else False)
resp.matched_blacklisted_images_rule = (
eval_result.bundle_decision.blacklisted_image.json()
if eval_result.bundle_decision.blacklisted_image else False)
resp.result = eval_result.as_table_json()
resp.created_at = int(time.time())
resp.evaluation_problems = [
problem_from_exception(i) for i in eval_result.errors
]
resp.evaluation_problems += [
problem_from_exception(i) for i in eval_result.warnings
]
if resp.evaluation_problems:
for i in resp.evaluation_problems:
log.warn(
"Returning evaluation response for image {}/{} w/tag {} and bundle {} that contains error: {}"
.format(user_id, image_id, tag, bundle["id"],
json.dumps(i.to_json())))
metrics.histogram_observe(
"anchore_policy_evaluation_time_seconds",
time.time() - timer,
status="fail",
)
else:
metrics.histogram_observe(
"anchore_policy_evaluation_time_seconds",
time.time() - timer,
status="success",
)
result = resp.to_json()
# Never let the cache block returning results
try:
if evaluation_cache_enabled and cache_mgr is not None:
cache_mgr.save(result)
except Exception as ex:
log.exception(
"Failed saving policy result in cache. Skipping and continuing."
)
db.commit()
return result
except HTTPException as e:
db.rollback()
log.exception("Caught exception in execution: {}".format(e))
raise
except Exception as e:
db.rollback()
log.exception("Failed processing bundle evaluation: {}".format(e))
return (
make_response_error(
"Unexpected internal error",
details={"message": str(e)},
in_httpcode=500,
),
500,
)
finally:
db.close()
def check_user_image_inline(user_id, image_id, tag, bundle):
"""
Execute a policy evaluation using the info in the request body including the bundle content
:param user_id:
:param image_id:
:param tag:
:param bundle:
:return:
"""
db = get_session()
try:
# Input validation
try:
img_obj = db.query(Image).get((image_id, user_id))
except:
abort(Response(response='Image not found', status=404))
if not img_obj:
log.info('Request for evaluation of image that cannot be found: user_id = {}, image_id = {}'.format(user_id, image_id))
abort(Response(response='Image not found', status=404))
# Build bundle exec.
problems = []
executable_bundle = None
try:
executable_bundle = build_bundle(bundle, for_tag=tag)
if executable_bundle.init_errors:
problems = executable_bundle.init_errors
except InitializationError as e:
log.exception('Bundle construction and initialization returned errors')
problems = e.causes
if not problems:
# Execute bundle
try:
eval_result = executable_bundle.execute(img_obj, tag, ExecutionContext(db_session=db, configuration={}))
except Exception as e:
log.exception('Error executing policy bundle {} against image {} w/tag {}: {}'.format(bundle['id'], image_id, tag, e.message))
abort(Response(response='Cannot execute given policy against the image due to errors executing the policy bundle: {}'.format(e.message), status=500))
else:
# Construct a failure eval with details on the errors and mappings to send to client
eval_result = build_empty_error_execution(img_obj, tag, executable_bundle, errors=problems, warnings=[])
if executable_bundle and executable_bundle.mapping and len(executable_bundle.mapping.mapping_rules) == 1:
eval_result.executed_mapping = executable_bundle.mapping.mapping_rules[0]
resp = PolicyEvaluation()
resp.user_id = user_id
resp.image_id = image_id
resp.tag = tag
resp.bundle = bundle
resp.matched_mapping_rule = eval_result.executed_mapping.json() if eval_result.executed_mapping else {}
resp.last_modified = int(time.time())
resp.final_action = eval_result.policy_decision.final_decision
resp.result = eval_result.as_table_json()
resp.created_at = int(time.time())
resp.evaluation_problems = [problem_from_exception(i) for i in eval_result.errors]
resp.evaluation_problems += [problem_from_exception(i) for i in eval_result.warnings]
if resp.evaluation_problems:
for i in resp.evaluation_problems:
log.warn('Returning evaluation response for image {}/{} w/tag {} and bundle {} that contains error: {}'.format(user_id, image_id, tag, bundle['id'], json.dumps(i.to_dict())))
return resp.to_dict()
except HTTPException as e:
db.rollback()
log.exception('Caught exception in execution: {}'.format(e))
raise
except Exception as e:
db.rollback()
log.exception('Failed processing bundle evaluation: {}'.format(e))
abort(Response('Unexpected internal error', 500))
finally:
db.close()
def check_user_image_inline(user_id, image_id, tag, bundle):
"""
Execute a policy evaluation using the info in the request body including the bundle content
:param user_id:
:param image_id:
:param tag:
:param bundle:
:return:
"""
timer = time.time()
db = get_session()
try:
# Input validation
if tag is None:
# set tag value to a value that only matches wildcards
tag = '*/*:*'
try:
img_obj = db.query(Image).get((image_id, user_id))
except:
abort(Response(response='Image not found', status=404))
if not img_obj:
log.info(
'Request for evaluation of image that cannot be found: user_id = {}, image_id = {}'
.format(user_id, image_id))
abort(Response(response='Image not found', status=404))
# Build bundle exec.
problems = []
executable_bundle = None
try:
# Allow deprecated gates here to support upgrade cases from old policy bundles.
executable_bundle = build_bundle(bundle,
for_tag=tag,
allow_deprecated=True)
if executable_bundle.init_errors:
problems = executable_bundle.init_errors
except InitializationError as e:
log.exception(
'Bundle construction and initialization returned errors')
problems = e.causes
eval_result = None
if not problems:
# Execute bundle
try:
eval_result = executable_bundle.execute(
img_obj, tag,
ExecutionContext(db_session=db, configuration={}))
except Exception as e:
log.exception(
'Error executing policy bundle {} against image {} w/tag {}: {}'
.format(bundle['id'], image_id, tag, e.message))
abort(
Response(
response=
'Cannot execute given policy against the image due to errors executing the policy bundle: {}'
.format(e.message),
status=500))
else:
# Construct a failure eval with details on the errors and mappings to send to client
eval_result = build_empty_error_execution(img_obj,
tag,
executable_bundle,
errors=problems,
warnings=[])
if executable_bundle and executable_bundle.mapping and len(
executable_bundle.mapping.mapping_rules) == 1:
eval_result.executed_mapping = executable_bundle.mapping.mapping_rules[
0]
resp = PolicyEvaluation()
resp.user_id = user_id
resp.image_id = image_id
resp.tag = tag
resp.bundle = bundle
resp.matched_mapping_rule = eval_result.executed_mapping.json(
) if eval_result.executed_mapping else False
resp.last_modified = int(time.time())
resp.final_action = eval_result.bundle_decision.final_decision.name
resp.final_action_reason = eval_result.bundle_decision.reason
resp.matched_whitelisted_images_rule = eval_result.bundle_decision.whitelisted_image.json(
) if eval_result.bundle_decision.whitelisted_image else False
resp.matched_blacklisted_images_rule = eval_result.bundle_decision.blacklisted_image.json(
) if eval_result.bundle_decision.blacklisted_image else False
resp.result = eval_result.as_table_json()
resp.created_at = int(time.time())
resp.evaluation_problems = [
problem_from_exception(i) for i in eval_result.errors
]
resp.evaluation_problems += [
problem_from_exception(i) for i in eval_result.warnings
]
if resp.evaluation_problems:
for i in resp.evaluation_problems:
log.warn(
'Returning evaluation response for image {}/{} w/tag {} and bundle {} that contains error: {}'
.format(user_id, image_id, tag, bundle['id'],
json.dumps(i.to_dict())))
anchore_engine.subsys.metrics.histogram_observe(
'anchore_policy_evaluation_time_seconds',
time.time() - timer,
status="fail")
else:
anchore_engine.subsys.metrics.histogram_observe(
'anchore_policy_evaluation_time_seconds',
time.time() - timer,
status="success")
return resp.to_dict()
except HTTPException as e:
db.rollback()
log.exception('Caught exception in execution: {}'.format(e))
raise
except Exception as e:
db.rollback()
log.exception('Failed processing bundle evaluation: {}'.format(e))
abort(Response('Unexpected internal error', 500))
finally:
db.close()