Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add a service check for Riak. #1187

Merged
merged 1 commit into from
Dec 18, 2014
Merged

Add a service check for Riak. #1187

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
46 changes: 12 additions & 34 deletions checks.d/riak.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
from httplib2 import Http, HttpLib2Error

class Riak(AgentCheck):
SERVICE_CHECK_NAME = 'riak.can_connect'

keys = [
"vnode_gets",
Expand Down Expand Up @@ -55,34 +56,29 @@ def __init__(self, name, init_config, agentConfig, instances=None):

self.prev_coord_redirs_total = -1


def check(self, instance):
url = instance['url']
default_timeout = self.init_config.get('default_timeout', 5)
timeout = float(instance.get('timeout', default_timeout))

aggregation_key = md5(url).hexdigest()
service_check_tags = ['url:%s' % url]

try:
h = Http(timeout=timeout)
resp, content = h.request(url, "GET")

except socket.timeout, e:
self.timeout_event(url, timeout, aggregation_key)
return

except socket.error, e:
self.timeout_event(url, timeout, aggregation_key)
return

except HttpLib2Error, e:
self.timeout_event(url, timeout, aggregation_key)
return
except (socket.timeout, socket.error, HttpLib2Error) as e:
self.service_check(self.SERVICE_CHECK_NAME, AgentCheck.CRITICAL,
message="Unable to fetch Riak stats: %s" % str(e),
tags=service_check_tags)

if resp.status != 200:
self.status_code_event(url, resp, aggregation_key)
self.service_check(self.SERVICE_CHECK_NAME, AgentCheck.CRITICAL,
tags=service_check_tags,
message="Unexpected status of %s when fetching Riak stats, " \
"response: %s" % (resp.status, content))

stats = json.loads(content)
self.service_check(
self.SERVICE_CHECK_NAME, AgentCheck.OK, tags=service_check_tags)

[self.gauge("riak." + k, stats[k]) for k in self.keys if k in stats]

Expand All @@ -92,21 +88,3 @@ def check(self, instance):
self.gauge('riak.coord_redirs', count)

self.prev_coord_redirs_total = coord_redirs_total

def timeout_event(self, url, timeout, aggregation_key):
self.event({
'timestamp': int(time.time()),
'event_type': 'riak_check',
'msg_title': 'riak check timeout',
'msg_text': '%s timed out after %s seconds.' % (url, timeout),
'aggregation_key': aggregation_key
})

def status_code_event(self, url, r, aggregation_key):
self.event({
'timestamp': int(time.time()),
'event_type': 'riak_check',
'msg_title': 'Invalid reponse code for riak check',
'msg_text': '%s returned a status of %s' % (url, r.status_code),
'aggregation_key': aggregation_key
})