Open Hub
Open Hub

 Settings |  Report Duplicate

1
I Use This!
Inactive

Commits : Listings

Analyzed 1 day ago. based on code collected 2 days ago.
Oct 13, 2024 — Oct 13, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
More clean with pylint
nap
More... about 16 years ago
Pylint pass for a clean code, more comments and aven a bug resolved in the cutting function (a previous h called instead of a service.host). Thanks Pylint.
nap
More... about 16 years ago
Now dependant checks will not rise a new check is the dependant elements (father for example) il already in check. It will be just add to the depend_on_me list of this check. This check will also be advanced so the dependancy revolation is quicker.
nap
More... about 16 years ago
Add a message when launching a freshness check
nap
More... about 16 years ago
Oups, new the var directory in the repository for test
nap
More... about 16 years ago
Add a new broker plugin : the service perfdata exporter. Pluins are great and usefull after all :)
nap
More... about 16 years ago
Now checks raise perf_data and schedulingitems get thems.
nap
More... about 16 years ago
Some changes for Perf: Less memory : inline services picke with getstate/setstate, due to slots. More Perf in Shinken : Timeperiod get_next_valid_time_from_t now have a cache. It speed up things a lot! Cache is clean One time a loop. More Perf in poller : put_results is now a oneway call. More perf on arbiter : no more self.has but hasattr instead :(
nap
More... about 16 years ago
Add compression to PYRO.
nap
More... about 16 years ago
Add a clean_queues function so if nobody ask us for checks, broks or actions, we can do not explose in memory. Max values are 2*number of elements. It's quite high. We can maybe add a drop_file to lose nothing.
nap
More... about 16 years ago
Comments on config cutting part.
nap
More... about 16 years ago
Big Perf boost in the config part : the cutting part is speedup a lot. Now we oad balance by host. Now servicedep are also --really-- take into account in the cutting part. The __slots__ of python are used for service for a big drop of emory usage. Still the brok that are very memory-vore.
nap
More... about 16 years ago
Solve a problem in the scheduling function... Less perf in fact... And the cut need to be changed...
nap
More... about 16 years ago
Clean of last print so We can test with a HUGE conf : and its a great result : event better than Nagios. 75000checks/5min with a latency of 2s in a Xeon5160. Diagram of architecture in doc is also updated. ..... 75000 :) :) :)
nap
More... about 16 years ago
Big speed boost : apply inheritance before explode, so less call to inherit. The has function is still call too often. Boost the iten class too, and change cutting functions. Theses functions need comments.
Jean
More... about 16 years ago
Add a remove_twins function to remove.. twins elemnts (services, hosts, contacts and timeperiods). Now we have the same number of service than a Nagios verify :)
Jean
More... about 16 years ago
Boost search by reversed list. And solved bug where service on hstgroup were created 2 times.
Jean
More... about 16 years ago
Clean some code for None problem. Add profile code and optimize cutting code (pygraph)
nap
More... about 16 years ago
Big scheduler clean. Now with an add function to manage add in queues like checks or actions, or brok, etc.
Jean
More... about 16 years ago
Now notifications raise notification_raise Brok. All data are not fill. I wait a ninja page to see what to put into :). notification code is clearer (was not very difficult...) and scheduler one. I'll clean scheduler more.
Jean
More... about 16 years ago
oups, a early commit... bad bash.
Jean
More... about 16 years ago
oups, a little code forgot...
Jean
More... about 16 years ago
oups, a little code forgot...
Jean
More... about 16 years ago
Clean the scheduler code for actions. It's need if we want to take broks from notifications. There are a other problem : the command is resolved when the notif is launched, but it may be wrong... The history is also important, so the command need to be resolved at notif creation, bat isn't it? :). So notification need to be fully rethink.... and then Broks :)
Jean
More... about 16 years ago
Clean brok code in host, hostgroup, service, servicegroup. Now the code is in item*py. A new class property is here my_type, a string that help create the brok 'initial_'service'_status' for example. Contact and contactgroup are now managed in brok too. But they are not used for the moment by ninja... but data in database are good.. I think... maybe... I hope... Now we can begin to manage notifications, comments, ackno and co. Log too? :)
Jean
More... about 16 years ago
Clean broker code. Now broker inherit from actionner, so more factorisez code, maybe we can do more. Bin (broker, poller, reactionner, shinken) have now a default port, so argv(1) is not mandatory. Need more beautiful arg option? Yes... And more code clean, in host/service.
Jean
More... about 16 years ago
Change Broker : now load plugins in plugins. They are in chagre of broks (so one brok -> N*brokers). I move the Mysql_merlin code into a plugin, and create a shell for ndo one. Need more clean (broker, and brok in host/services)
nap
More... about 16 years ago
Begining of the Brok clean
Jean
More... about 16 years ago
Add status brok information for hostgroup and servicegroup, change id of internal object so 0 is no more used (pb with auto_increment). Still need clean in the broker and brok code, it's quite ugly...
Jean
More... about 16 years ago
Add brok update for host/service so the status is now OK in database when check reeturns. External command raise brok too (global conf update-> brok for conf). code need clean and factorised...
nap
More... about 16 years ago