commit eeae33fbbacf4de14d260ba24a1db0ae2e048538 Author: Damian Johnson atagar@torproject.org Date: Wed Aug 6 08:25:29 2014 -0700

Defaulting suppression time based on issue severity

Frequently I tone down our NOTICE issues to have a bit of suppression to cut down on noise for the list. Just realized there's no reason not to make this a blanket policy - defaulting suppression duration based on the runlevel, with it being overwritten in our config. --- consensus_health_checker.py | 9 ++++++++- data/consensus_health.cfg | 12 ++---------- 2 files changed, 10 insertions(+), 11 deletions(-)

diff --git a/consensus_health_checker.py b/consensus_health_checker.py index e85ece1..2ce14a4 100755 --- a/consensus_health_checker.py +++ b/consensus_health_checker.py @@ -102,7 +102,14 @@ class Issue(object): except ValueError: log.error("Non-numic suppression time (%s): %s" % (self._template, suppression_duration))

- return 0 + # Default to suppression based on the severity of the issue. + + if self.get_runlevel() == Runlevel.NOTICE: + return 24 # 1 day + elif self.get_runlevel() == Runlevel.WARN: + return 4 # 4 hours + else: + return 0 # no suppression for errors

def __str__(self): return "%s: %s" % (self.get_runlevel(), self.get_message()) diff --git a/data/consensus_health.cfg b/data/consensus_health.cfg index 2201a57..ceb54ed 100644 --- a/data/consensus_health.cfg +++ b/data/consensus_health.cfg @@ -20,17 +20,9 @@ msg AUTHORITY_UNAVAILABLE => Unable to retrieve the {fetch_type} from {authority

# hours that we'll suppress messages if it hasn't changed

-suppression DIFFERENT_RECOMMENDED_VERSION => 24 # 1 day -suppression UNKNOWN_CONSENSUS_PARAMETERS => 24 # 1 day -suppression MISMATCH_CONSENSUS_PARAMETERS => 24 # 1 day suppression CERTIFICATE_ABOUT_TO_EXPIRE => 336 # 2 weeks -suppression MISSING_BANDWIDTH_SCANNERS => 24 # 1 day -suppression EXTRA_BANDWIDTH_SCANNERS => 24 # 1 day -suppression MISSING_AUTHORITIES => 168 # 1 week -suppression EXTRA_AUTHORITIES => 168 # 1 week -suppression TOR_OUT_OF_DATE => 168 # 1 week -suppression BADEXIT_OUT_OF_SYNC => 24 # 1 day -suppression BANDWIDTH_AUTHORITIES_OUT_OF_SYNC => 24 # 1 day +suppression MISSING_AUTHORITIES => 24 # 1 day +suppression TOR_OUT_OF_DATE => 24 # 1 day suppression AUTHORITY_UNAVAILABLE => 24 # 1 day

# directory authorities for which to not provide notifications