I am trying to understand collectd notifications but could not understand how exactly it works.
What plugins raise the notifications, what plugin receives the notifications and how are they handled? And how these notifications can be integrated in system like nagios and/or graphite.
I tried to google any sample depicting usage, but could not find one.
This is quite well explained on the wiki: https://collectd.org/wiki/index.php/Notifications_and_thresholds.
In short words: some plugins produce notifications, e.g. the 'df' plugin, and some consume them like the 'syslog' plugin. There's currently no "nagios consumer", but you can use the shipped 'collectd-nagios' helper as an interface to nagios using 'nrpe'.
Related
Is there a way to view a list of all the files that people have uploaded/attached to a channel? Perhaps even through the search feature or a slash command? Or maybe there's a simple plugin that adds this functionality?
As a comparison, this is a basic feature of HipChat and Slack, but I don't see it in Mattermost. I'm trying out Mattermost Team Edition version 3.6.2. But for the most part, Mattermost is a great and most welcome app.
Now there is a plugin for that, compatible with versions 5.12+:
https://github.com/Amonith/mattermost-file-list
Disclaimer: I'm the author.
FWIW, meanwhile there is a comprehensive UI available from the "Search" edit to allow just that:
Mattermost currently (version 3.6) does not have any feature or API endpoint to allow listing all attachments in a channel, nor any way of searching for messages with attachments, so this currently isn't possible.
Worklight 6.2.0
Native Worklight App on Samsung Galaxy S4, Android 4.4.2
WLAnalytics.enable();
WLAnalytics.log("some text", new org.json.JSONOBject() );
WLAnalytics.send();
// and also go on to successfully call an adapter
Analytics Dasboard shows the app version and adapter activity. Log Search does not show any application log messages and the dropdown for selecting applications shows "All Applications" only, no sign of my app.
Have I missed some initialisation step? Any other ideas?
** edited to add **
It has been suggested that we should use the method:
WLAnalytics.log("some text");
In our 6.2.0.00 CLI environment there is no such Java method.
The answer is that there a further initialisation requirement that seems to be necessary when working with a pure Native application, these are typically build using the Worklight CLI tooling.
This is the initialisation, note the call to Logger.setContext()
WLAnalytics.enable();
Logger.setContext(this);
Then this works
WLAnalytics.log("My test message2", new org.json.JSONObject());
It's worth noting that the call to WLAnalytics.send() is not necessary in normal running as typically the analytic data is buffered and sent as a piggy-back on adapter calls. However while testing a call to send() does help.
Further, if running in an environment where the Analytics WAR is on a separate machine from the Worklight Server WAR there are additional latencies. Hence testing all of this needs care.
For now, I suggest that you just use the WLAnalytics.log(String) method. There are some clear inconsistencies that need to be dealt with whether it be through documentation or code fixes.
The WL.Logger APIs were originally created to send log data to a custom adapter, which is why they take a dictionary/object for extra metadata. The data sent to the custom adapter could be read as a valid JSON object to run operations on the adapter.
The WL.Analytics APIs mimicked the WL.Logger APIs for the same purpose: parsing the JSON on a worklight adapter. The Operational analytics server came as a convenience to intercept and display some of these logs, but not all of them are being captured as you have learned.
Your questions are all valid though, as none of this is described in the documentation. In future releases, we may make use of the extra JSON object passed into the API in the Operational analytics console, but for right now they only serve their original purpose of sending the analytics to a custom adapter.
I would like to receive remote push notification when the app is on the foreground. I think I can use "didReceiveRemoteNotification" in Objective-C, but how about in Titanium? I've been searching solutions but can't find anything. Does anybody help me out?
Titanium currently only supports local notifications (by default) I believe. You might want to take a look at AppCelerator as a solution.
Sorry, i was wrong last time. It's very hard to find but Titanium supports a way of interacting with PushNotifications.
Please take a look at Ti.Network.registerForPushNotifications. It provides a callback that is executed when you receive a Push Notification.
The other posted solutions below may provide a better integration for titanium so it's up to you what you want to use. But this should answer your original question.
Unfortunately Titanium only supports local notifications on iOS. If
you want to support remote notifications you can use
Urban Airship
ACS (Appcelerator Cloud Services) or
you write your own module that provides "didReceiveRemoteNotifcation" as an app event to your app
As far as i know you can only listen on local notifications by
default but you can try this method if it reacts on remote
notifications..
Even though the API has been open since Mac OS X Leopard, there's surprisingly, and unfortunately, very little documentation on how to correctly use SMJobBless() for creating privileged helper tools. There are a lot of gotchas, even when copying code directly from Apple's sample project. Luckily, I've found my way around this, and have gotten the basis for my helper tool working.
However, it would seem that SMJobBless() only blesses the tool and copies it over, but doesn't run it. I've included code in my helper tool's main() function that should run, but doesn't (since NSLog() inexplicably doesn't work–according to the tiny bit of information I have found–I've tried syslog()ing some "Hello world" type strings, but nothing appears on the system console). There's no indication that the helper tool is launched at all.
The documentation is mostly useless. It simply says that after SMJobBless() is called, the helper tool is 'ready', with no indication of what 'ready' even means.
Furthermore, Apple's sample doesn't include any interprocess communication code, and doesn't explain how one is supposed to interact with the helper tool. Do you use Distributed Objects? Mach ports? Who knows? There's no official word on how to do it.
So, does anyone have any information on how to get this done? I've confirmed that the helper tool is installed, and authentication works, but I simply can't figure out how to launch the helper tool and communicate with it - there's simply such a gap in the documentation that this is a mystery for now. It's very frustrating; I can't be the only one with this problem (but there's little mention of it anywhere), and SMJobBless() obviously works somehow, since it's what Apple uses.
(Please don't mention AuthorizationExecuteWithPrivileges(). I'm not using it: it's deprecated, sure to go away, and is a major security hole. No thanks.)
XPC isn't an option if you're trying to elevate privileges (from https://developer.apple.com/library/archive/documentation/MacOSX/Conceptual/BPSystemStartup/Chapters/CreatingXPCServices.html):
By default, XPC services are run in the most restricted environment
possible—sandboxed with minimal filesystem access, network access, and
so on. Elevating a service’s privileges to root is not supported.
SMJobBless will install a helper tool and register it with Launchd, as in the SMJobBless example provided by Apple. The trick to getting your helper tool to actually launch is to simply attempt to connect to your helper tool's advertised services.
There was a WWDC2010 example called ssd that demonstrated a simple launchd client/server model via sockets. It's not available from Apple any longer, but I've found a link here: https://lists.apple.com/archives/macnetworkprog/2011/Jul/msg00005.html
I've incorporated the dispatch queue handling in the server code from the ssd example into the helper tool in the SMJobBless example and can confirm that my helper tool is indeed running (as root) when my main app attempts a connection on the appropriate port. See the WWDC2010 video on Launchd to understand the other mechanisms with which you can communicate with your helper tool (other than sockets).
I'm not sure I can legally redistribute the modified sources I have, but it should be fairly straightforward to merge the two projects and get your helper tool running.
Edit: Here is an example project I wrote that uses a distributed object for communication between the app and helper: https://www.dropbox.com/s/5kjl8koyqzvszrl/Elevator.zip
In fact #KurtRevis's comment is right, you can use XPC APIs without using XPC services, and it is ideally suited to the job since.
Nathan de Vries has an excellent writeup of using XPC APIs with SMJobBless and has even modified the SMJobBless sample app to use mach XPC to both activate the job and for bidirectional communications:
http://atnan.com/blog/2012/02/29/modern-privileged-helper-tools-using-smjobbless-plus-xpc/
https://github.com/atnan/SMJobBlessXPC
Somewhat related to all this is avoiding unnecessary admin password prompts. See the following email list thread for ideas on how to check if the bundle version and code signature of an already installed helper match (which also allows you to remove a higher versioned helper in the case of a user downgrading):
http://www.cocoabuilder.com/archive/cocoa/309298-question-about-smjobbless.html
If you don't want to wade through the thread, here is a link to the modified SMJobBless sample project provided by Eric Gorr:
http://ericgorr.net/cocoadev/SMJobBless.zip
Also note that the ssd example mentioned in other answers here is still available online from Apple as part of the WWDC 2010 download bundle:
http://connect.apple.com/cgi-bin/WebObjects/MemberSite.woa/wa/getSoftware?code=y&source=x&bundleID=20645
Apple now (2015) has an "EvenBetterAuthorizationSample" that demonstrates installing a privileged helper tool and using the NSXPCConnection API to communicate between the app and the helper tool:
https://developer.apple.com/library/mac/samplecode/EvenBetterAuthorizationSample/Listings/Read_Me_About_EvenBetterAuthorizationSample_txt.html
The README is some of the best (only?) documentation of SMJobBless() available.
I feel your pain and am in the same boat. I'm in charge of the Mac version of an app that needs to perform various system configuration tasks. Of course some of these task need to be done with administrative rights. I started by using the sample code from BetterAuthorizationSample. It was a major pain to implement but it seemed to work. But then ran into cases where it would crash on some systems. I didn't understand everything that the BAS code did and my own lack of coding experience probably contributed to the problems. So I had to remove these privileged functions from my app.
Apple doesn't seem to care about the lack of documentation. See this message from the creator of the ServiceManagement framework. From his comments, I assume that XPC is the "intuitive replacement" he is referring to, but since it is only available on Lion, you'll still have to find another solution for Snow Leopard or earlier clients. It's also not clear to me if XPC can be used for privileged helpers (system level tasks that require admin or root access) or is just intended for privilege separation within your own app to make it more secure.
The BAS documentation is in desperate need of an update, but it also doesn't appear to be a top priority.
Now I'm attempting to rewrite my app from the ground up. Professional Cocoa Application Security by Graham Lee gives some insight on how to do use privileged helpers with SMJobBless, but doesn't go into much detail about on-demand access to launchd jobs.
So here's what I've been able to find:
If you want to launch your privileged helper on demand, you'll have to use an IPC socket. You should add a Sockets entry to your helper's launchd.plist. After you install the app with SMJobBless, the helper will need to "check-in" with launchd (via LAUNCH_KEY_CHECKIN) to get the socket file descriptors.
Sadly, the only mentions of LAUNCH_KEY_CHECKIN seem to be in the SampleD and BAS sample code.
I don't have any experience with sockets, so that's my roadblock at the moment. I'd like to use the highest level API I can, so I'm trying to find out if I can use any Objective-C classes for this (like NSStream).
You might find the launchd developers mailing list helpful. Another XPC option I just found out about is XPCKit. It's worth a look.
HTH
I wrote a blog post on this a few months ago, which included a cleaned up version of Apple's SMJobBless sample. Might help...
http://www.bornsleepy.com/bornsleepy/os-x-helper-applications
Itai have you looked at the SMJobBless sample code from WWDC 2010? It includes a helper tool and app to bless it.
https://developer.apple.com/library/mac/#samplecode/SMJobBless/Listings/ReadMe_txt.html#//apple_ref/doc/uid/DTS40010071-ReadMe_txt-DontLinkElementID_3
Its README file says:
This sample as it stands does not actually run the helper tool. The following samples show how to [sic] a launchd job and set up interprocess communication:
ssd (Doesn't seem to be online anymore. Was part of the WWDC 2010 sample code.)
BetterAuthorizationSample
In Zabbix, there is a well documented (although in Draft status) API for instantiating targets and controlling them. Does such an API exist in Nagios?
The need for this is a cloud-based implementation. I would like to be able to scale monitoring targets dynamically so that on Monday it would be monitoring 10 targets and Tuesday it would be monitoring 15 and down to 8 on Wednesday.
The best way I can figure to do this is to control targets in a programmatic fashion since auto-discovery will probably fail as the nodes are not in the same subnet. Any ideas on how to do this with Nagios?
Nagios does not natively have this ability. I have gotten around this in the past by adding scripts to contact the nagios server and register on first boot and then using nagios to destroy servers either automatically or via user command. These are custom scripts though.
I have also passed through this problem and yes Nagios does not have any kind of Command Line API but I have discovered that NagiosQL has this feature. http://www.nagiosql.org/
There are many advantages for using such a tool but I will not touch this here...
Regarding the command line API, NagiosQL can manage multiple domains through web interface or command line api. Im my opinion this is very useful.