-
I've added a new topic related to zc.buildout and SimpleMonitor in the wiki SimpleMonitor is a lightweight monitor tool that can be used to watch a buildout instance and send alerts in case something goes down (like a zope instance) It acts a bit like Nagios but can work under windows. This means that it would be easy and very useful to create a recipe to be able to set up monitoring features in a buildout http://jamesoff.net/site/code/simplemonitor/ I have contacted the author and he will be delighted to see his tool working under python 2.4 (it uses 2.5 statements at this time) If you are interested, mark your name in this topic ! ;) Tarek -- Tarek Ziadé - Directeur Technique INGENIWEB (TM) - SAS 50000 Euros - RC B 438 725 632 Bureaux de la Colline - 1 rue Royale - Bâtiment D - 9ème étage 92210 Saint Cloud - France Phone : 01.78.15.24.00 / Fax : 01 46 02 44 04 http://www.ingeniweb.com - une société du groupe Alter Way
- Thread Outline:
-
- Re: New topic ! SimpleMonitor recipe fo ... by "Sylvain Viollon" <thefunny@...>
-
Hello, On Wed, Apr 23, 2008 at 5:34 PM, Tarek Ziade <tarek.ziade@...> wrote: > I've added a new topic related to zc.buildout and SimpleMonitor in the wiki > > SimpleMonitor is a lightweight monitor tool that can be used to watch a > buildout instance and send alerts > in case something goes down (like a zope instance) > > It acts a bit like Nagios but can work under windows. This means that it > would be easy and very useful to > create a recipe to be able to set up monitoring features in a buildout > Nagios works under windows as well. [...] Sylvain, -- Qui sème le vent récolte l'ouragan
-
2008/4/23 Sylvain Viollon <thefunny@...>: > Hello, > > On Wed, Apr 23, 2008 at 5:34 PM, Tarek Ziade <tarek.ziade@...> > wrote: > > I've added a new topic related to zc.buildout and SimpleMonitor in the > wiki > > > > SimpleMonitor is a lightweight monitor tool that can be used to watch a > > buildout instance and send alerts > > in case something goes down (like a zope instance) > > > > It acts a bit like Nagios but can work under windows. This means that it > > would be easy and very useful to > > create a recipe to be able to set up monitoring features in a buildout > > > > Nagios works under windows as well. as far as we know, you cannot run the whole system under Windows, http://www.nagios.org/about/ but if you know how, this is interesting ! > > [...] > > Sylvain, > > -- > Qui sème le vent récolte l'ouragan > > > -- > Archive: > http://www.openplans.org/projects/plone-3-paris-sprint/lists/plone-3-paris-sprint-discussion/archive/2008/04/1208967849082 > To unsubscribe send an email with subject unsubscribe to > plone-3-paris-sprint-discussion@.... Please contact > plone-3-paris-sprint-discussion-manager@... for questions. > > -- Tarek Ziadé - Directeur Technique INGENIWEB (TM) - SAS 50000 Euros - RC B 438 725 632 Bureaux de la Colline - 1 rue Royale - Bâtiment D - 9ème étage 92210 Saint Cloud - France Phone : 01.78.15.24.00 / Fax : 01 46 02 44 04 http://www.ingeniweb.com - une société du groupe Alter Way
-
On Wed, 23 Apr 2008 19:44:49 +0200 "Tarek Ziade" <tarek.ziade@...> wrote: > 2008/4/23 Sylvain Viollon <thefunny@...>: > > > Hello, > > > > On Wed, Apr 23, 2008 at 5:34 PM, Tarek Ziade > > <tarek.ziade@...> wrote: > > > I've added a new topic related to zc.buildout and SimpleMonitor > > > in the > > wiki > > > > > > SimpleMonitor is a lightweight monitor tool that can be used to > > > watch a buildout instance and send alerts > > > in case something goes down (like a zope instance) > > > > > > It acts a bit like Nagios but can work under windows. This means > > > that it would be easy and very useful to > > > create a recipe to be able to set up monitoring features in a > > > buildout > > > > > > > Nagios works under windows as well. > > > as far as we know, you cannot run the whole system under Windows, > > http://www.nagios.org/about/ > > but if you know how, this is interesting ! > > Well, Nagios 2 is regular C-Unix (without funky stuff) and Perl. Both work well under windows with cygwin. If you search through Nagios mailing list, you will see that several users report this setup. I don't think that Nagios want to advertise about it, because it's maintained by a company which doesn't want to make support to windows user^Wadministrator which just known how to do next next (it's can be tricky to set up and manage, since it's an hack). After, I don't think you want a monitoring per buildout. A monitoring service is a service that let you to be sure that's all your others services are running correctly. If you have a server with 10 applications using this buildout setup, how are you going to be sure that these 10 buildout monitoring systems works correctly ? It's better to have only one service/setup which check everything, and here you will have to check only that this one works (and in the same time, you will see if others services fails). Or if you're not sure you can setup an redundant server with Nagios as well. Unless you have an application which really doesn't work, and here you can put a zopectl restart every hour in a crontab. Sylvain, -- Sylvain Viollon -- Infrae t +31 10 243 7051 -- http://infrae.com Hoevestraat 10 3033GC Rotterdam -- The Netherlands
-
2008/4/24 Sylvain Viollon <sylvain@...>: > On Wed, 23 Apr 2008 19:44:49 +0200 > "Tarek Ziade" <tarek.ziade@...> wrote: > > > 2008/4/23 Sylvain Viollon <thefunny@...>: > > > > > Hello, > > > > > > On Wed, Apr 23, 2008 at 5:34 PM, Tarek Ziade > > > <tarek.ziade@...> wrote: > > > > I've added a new topic related to zc.buildout and SimpleMonitor > > > > in the > > > wiki > > > > > > > > SimpleMonitor is a lightweight monitor tool that can be used to > > > > watch a buildout instance and send alerts > > > > in case something goes down (like a zope instance) > > > > > > > > It acts a bit like Nagios but can work under windows. This means > > > > that it would be easy and very useful to > > > > create a recipe to be able to set up monitoring features in a > > > > buildout > > > > > > > > > > Nagios works under windows as well. > > > > > > as far as we know, you cannot run the whole system under Windows, > > > > http://www.nagios.org/about/ > > > > but if you know how, this is interesting ! > > > > > > Well, Nagios 2 is regular C-Unix (without funky stuff) and Perl. Both > work well under windows with cygwin. If you search through Nagios > mailing list, you will see that several users report this setup. > > I don't think that Nagios want to advertise about it, because it's > maintained by a company which doesn't want to make support to windows > user^Wadministrator which just known how to do next next (it's can be > tricky to set up and manage, since it's an hack). Oh OK, good to know thanks. I feel like SimpleMonitor is simpler to hack around and adapt in that case, because it is plain Python, but that is just my opinion. > > > > After, I don't think you want a monitoring per buildout. A monitoring > service is a service that let you to be sure that's all your others > services are running correctly. If you have a server with 10 > applications using this buildout setup, how are you going to be sure > that these 10 buildout monitoring systems works correctly ? It's better > to have only one service/setup which check everything, and here you will > have to check only that this one works (and in the same time, you > will see if others services fails). Or if you're not sure you can setup > an redundant server with Nagios as well. > zc.buildout can be used for applications setups and is not plone/zope centric. For instance, our buildbot is a buildout. By having a buildout recipe that allows setting a monitor in a matter of lines, let's say: [monitor] recipe = collective.recipe.simplemonitor watch = 8080 TCP mail ... 8282 TCP ... ... This can be used either in a zope buildout on servers that have only one zope server, either on a buildout dedicated to monitor several applications (and to do other things maybe) > Unless you have an application which really doesn't work, and here > you can put a zopectl restart every hour in a crontab. zopectl restart is not a solution imho, because when a zope server drop, even if it is ok to restart it, the problem that put it down should be corrected. ++ Tarek > > Sylvain, > > > > -- > Sylvain Viollon -- Infrae > t +31 10 243 7051 -- http://infrae.com > Hoevestraat 10 3033GC Rotterdam -- The Netherlands > -- Tarek Ziadé - Directeur Technique INGENIWEB (TM) - SAS 50000 Euros - RC B 438 725 632 Bureaux de la Colline - 1 rue Royale - Bâtiment D - 9ème étage 92210 Saint Cloud - France Phone : 01.78.15.24.00 / Fax : 01 46 02 44 04 http://www.ingeniweb.com - une société du groupe Alter Way
-
On Thu, 24 Apr 2008 10:24:10 +0200 "Tarek Ziade" <tarek.ziade@...> wrote: > 2008/4/24 Sylvain Viollon <sylvain@...>: > > > On Wed, 23 Apr 2008 19:44:49 +0200 > > "Tarek Ziade" <tarek.ziade@...> wrote: > > > > > 2008/4/23 Sylvain Viollon <thefunny@...>: > > > > > > > Hello, > > > > > > > > On Wed, Apr 23, 2008 at 5:34 PM, Tarek Ziade > > > > <tarek.ziade@...> wrote: > > > > > I've added a new topic related to zc.buildout and > > > > > SimpleMonitor in the > > > > wiki > > > > > > > > > > SimpleMonitor is a lightweight monitor tool that can be used > > > > > to watch a buildout instance and send alerts > > > > > in case something goes down (like a zope instance) > > > > > > > > > > It acts a bit like Nagios but can work under windows. This > > > > > means that it would be easy and very useful to > > > > > create a recipe to be able to set up monitoring features in a > > > > > buildout > > > > > > > > > > > > > Nagios works under windows as well. > > > > > > > > > as far as we know, you cannot run the whole system under Windows, > > > > > > http://www.nagios.org/about/ > > > > > > but if you know how, this is interesting ! > > > > > > > > > > Well, Nagios 2 is regular C-Unix (without funky stuff) and Perl. > > Both work well under windows with cygwin. If you search through > > Nagios mailing list, you will see that several users report this > > setup. > > > > I don't think that Nagios want to advertise about it, because it's > > maintained by a company which doesn't want to make support to > > windows user^Wadministrator which just known how to do next next > > (it's can be tricky to set up and manage, since it's an hack). > > > Oh OK, good to know thanks. I feel like SimpleMonitor is simpler to > hack around and adapt in that case, > because it is plain Python, but that is just my opinion. > > > > > > > > > After, I don't think you want a monitoring per buildout. A > > monitoring service is a service that let you to be sure that's all > > your others services are running correctly. If you have a server > > with 10 applications using this buildout setup, how are you going > > to be sure that these 10 buildout monitoring systems works > > correctly ? It's better to have only one service/setup which check > > everything, and here you will have to check only that this one > > works (and in the same time, you will see if others services > > fails). Or if you're not sure you can setup an redundant server > > with Nagios as well. > > > > zc.buildout can be used for applications setups and is not plone/zope > centric. > For instance, our buildbot is a buildout. > Yes I known that. But I found silly to install 10 same things on a computer then you need to have only one, even if it's easier. For instance, I found silly as well to install apache in your buildout, and not use the system one. You don't have anymore update and security update, unless you do it by yourself, and your setup have been check over other security issue which may happens due to it. But, pay attention, I don't see the advantages to *install it*, but you can still provides an configuration file to *run* apache if you want in your buildout (assuming you have only one application on your computer, which is 80% never the case). > By having a buildout recipe that allows setting a monitor in a matter > of lines, let's say: > > [monitor] > recipe = collective.recipe.simplemonitor > > watch = > 8080 TCP mail ... > 8282 TCP > ... > > ... > > > This can be used either in a zope buildout on servers that have only > one zope server, > either on a buildout dedicated to monitor several applications (and > to do other things maybe) > If you make an buildout only for monitoring I agree, which monitor others buildout. But what's the advantages over a existing system, which works, is tested in real condition (since more people use it), and where you can found a lot more *administrator* to manage it ? (Even if it's easy, people usually don't want to make effort to learn new stuff when they known a working way). > > > Unless you have an application which really doesn't work, and here > > you can put a zopectl restart every hour in a crontab. > > > zopectl restart is not a solution imho, because when a zope server > drop, even if it is ok to restart > it, the problem that put it down should be corrected. > > I agree on this point, but sometimes it's difficult to dig it, and it's the temporary solution (which can last years by the way). Sylvain, -- Sylvain Viollon -- Infrae t +31 10 243 7051 -- http://infrae.com Hoevestraat 10 3033GC Rotterdam -- The Netherlands
-
2008/4/24 Sylvain Viollon <sylvain@...>: > cut > > > > zc.buildout can be used for applications setups and is not plone/zope > > centric. > > For instance, our buildbot is a buildout. > > > > Yes I known that. But I found silly to install 10 same things on a > computer then you need to have only one, even if it's easier. For > instance, I found silly as well to install apache in your buildout, > and not use the system one. > You don't have anymore update and security > update, unless you do it by yourself, and your setup have been check > over other security issue which may happens due to it. > Who said so ? I am just talking about make it easier to configure softwares, not recreate a system in your buildot. In some case though, it is not silly at all to install some parts in a buildout For instance, buildbot is a simple egg, so an easy_install is simpler that using a system-wide installation. But this is a topic that is currently discussed at python level, since setuptools will be the next distutils and since eggs where blessed by Guido. This is a tough problem anyway, wheter the packages upgrades are system package-centric or application-centric. If all your customers are under debian or red-hat, it makes a lot of sense to use debs and rmps. But when you have some customers under windows, a standardized way to distribute your app is not obvious at all. I'd be happy to talk about that at the sprint, to compare our distribution models > > But, pay attention, I don't see the advantages to *install it*, but > you can still provides an configuration file to *run* apache if you > want in your buildout (assuming you have only one application on your > computer, which is 80% never the case). > Yes, exactly, and this is the same thing for monitoring needs imho > > > By having a buildout recipe that allows setting a monitor in a matter > > of lines, let's say: > > > > [monitor] > > recipe = collective.recipe.simplemonitor > > > > watch = > > 8080 TCP mail ... > > 8282 TCP > > ... > > > > ... > > > > > > This can be used either in a zope buildout on servers that have only > > one zope server, > > either on a buildout dedicated to monitor several applications (and > > to do other things maybe) > > > If you make an buildout only for monitoring I agree, which monitor > others buildout. > Whatever fits your need. > But what's the advantages over a existing system, which > works, is tested in real condition (since more people use it), and where > you can found a lot more *administrator* to manage it ? (Even if it's > easy, people usually don't want to make effort to learn new stuff when > they known a working way). > The big advantage is to automatically provide simple monitoring services when you deliver, let's say, a plone application. We have for instance a recipe to set up a log rotation (rotatzlog) The real blast is to be able to define monitoring in a few configuration lines, and let the system install (or not) then configure the damn thing, no matter what the platform is. I am not debating on what is the best solution, if you have a Nagios administrator available, fine, it would be dumb to provide such a thing, and not use his work. But if you deliver a application to a customer that don't even have an administrator, being able to provide such tools makes a lot of sense. If it works by itself, on every platform, without having to install a lot of third-party software, that's even better. > > > > > Unless you have an application which really doesn't work, and here > > > you can put a zopectl restart every hour in a crontab. > > > > > > zopectl restart is not a solution imho, because when a zope server > > drop, even if it is ok to restart > > it, the problem that put it down should be corrected. > > > > > > I agree on this point, but sometimes it's difficult to dig it, and > it's the temporary solution (which can last years by the way). > Sure. Anyways, we do need such a recipe for our needs ;) so maybe the sprint is not the right place to work on this, I don't know.. unless someone else feels he has the same kinf of need. Regards Tarek > > Sylvain, > > > -- > Sylvain Viollon -- Infrae > t +31 10 243 7051 -- http://infrae.com > Hoevestraat 10 3033GC Rotterdam -- The Netherlands > -- Tarek Ziadé - Directeur Technique INGENIWEB (TM) - SAS 50000 Euros - RC B 438 725 632 Bureaux de la Colline - 1 rue Royale - Bâtiment D - 9ème étage 92210 Saint Cloud - France Phone : 01.78.15.24.00 / Fax : 01 46 02 44 04 http://www.ingeniweb.com - une société du groupe Alter Way
-
-
-
-
-
text.html (text/html) 1.1 kB