summaryrefslogtreecommitdiffstats
path: root/doc
diff options
context:
space:
mode:
Diffstat (limited to 'doc')
-rw-r--r--doc/README3
-rw-r--r--doc/developer-guidelines.html931
-rw-r--r--doc/developer-guidelines.sgml483
3 files changed, 1417 insertions, 0 deletions
diff --git a/doc/README b/doc/README
new file mode 100644
index 00000000..388bc1d7
--- /dev/null
+++ b/doc/README
@@ -0,0 +1,3 @@
1The developer documentation here is generated from the DocBook format.
2
3
diff --git a/doc/developer-guidelines.html b/doc/developer-guidelines.html
new file mode 100644
index 00000000..efac605f
--- /dev/null
+++ b/doc/developer-guidelines.html
@@ -0,0 +1,931 @@
1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN">
2<HTML
3><HEAD
4><TITLE
5>Nagios plug-in development guidelines</TITLE
6><META
7NAME="GENERATOR"
8CONTENT="Modular DocBook HTML Stylesheet Version 1.64
9"></HEAD
10><BODY
11CLASS="BOOK"
12BGCOLOR="#FFFFFF"
13TEXT="#000000"
14LINK="#0000FF"
15VLINK="#840084"
16ALINK="#0000FF"
17><DIV
18CLASS="BOOK"
19><A
20NAME="AEN1"
21></A
22><DIV
23CLASS="TITLEPAGE"
24><H1
25CLASS="TITLE"
26><A
27NAME="AEN3"
28>Nagios plug-in development guidelines</A
29></H1
30><H3
31CLASS="AUTHOR"
32><A
33NAME="AEN5"
34>Karl DeBisschop</A
35></H3
36><DIV
37CLASS="AFFILIATION"
38><DIV
39CLASS="ADDRESS"
40><P
41CLASS="ADDRESS"
42>karl@debisschop.net</P
43></DIV
44></DIV
45><H3
46CLASS="AUTHOR"
47><A
48NAME="AEN11"
49>Ethan Galstad</A
50></H3
51><DIV
52CLASS="AFFILIATION"
53><DIV
54CLASS="ADDRESS"
55><P
56CLASS="ADDRESS"
57>netsaint@linuxbox.com</P
58></DIV
59></DIV
60><H3
61CLASS="AUTHOR"
62><A
63NAME="AEN21"
64>Hugo Gayosso</A
65></H3
66><DIV
67CLASS="AFFILIATION"
68><DIV
69CLASS="ADDRESS"
70><P
71CLASS="ADDRESS"
72>hgayosso@gnu.org</P
73></DIV
74></DIV
75><H3
76CLASS="AUTHOR"
77><A
78NAME="AEN27"
79>Subhendu Ghosh</A
80></H3
81><DIV
82CLASS="AFFILIATION"
83><DIV
84CLASS="ADDRESS"
85><P
86CLASS="ADDRESS"
87>sghosh@sourceforge.net</P
88></DIV
89></DIV
90><H3
91CLASS="AUTHOR"
92><A
93NAME="AEN33"
94>Stanley Hopcroft</A
95></H3
96><DIV
97CLASS="AFFILIATION"
98><DIV
99CLASS="ADDRESS"
100><P
101CLASS="ADDRESS"
102>stanleyhopcroft@sourceforge.net</P
103></DIV
104></DIV
105><P
106CLASS="COPYRIGHT"
107>Copyright &copy; 2000 2001 2002 by Karl DeBisschop, Ethan Galstad,
108 Hugo Gayosso, Stanley Hopcroft, Subhendu Ghosh</P
109><HR></DIV
110><DIV
111CLASS="TOC"
112><DL
113><DT
114><B
115>Table of Contents</B
116></DT
117><DT
118><A
119HREF="#PREFACE"
120>About the guidelines</A
121></DT
122><DD
123><DL
124><DT
125><A
126HREF="#AEN51"
127>Copyright</A
128></DT
129></DL
130></DD
131><DT
132><A
133HREF="#AEN56"
134></A
135></DT
136><DD
137><DL
138><DT
139><A
140HREF="#PLUGOUTPUT"
141>Plugin Output for Nagios</A
142></DT
143><DD
144><DL
145><DT
146><A
147HREF="#AEN60"
148>Print only one line of text</A
149></DT
150><DT
151><A
152HREF="#AEN63"
153>Screen Output</A
154></DT
155><DT
156><A
157HREF="#AEN67"
158>Return the proper status code</A
159></DT
160><DT
161><A
162HREF="#AEN71"
163>Plugin Return Codes</A
164></DT
165></DL
166></DD
167><DT
168><A
169HREF="#SYSCMDAUXFILES"
170>System Commands and Auxiliary Files</A
171></DT
172><DD
173><DL
174><DT
175><A
176HREF="#AEN117"
177>Don't execute system commands without specifying their
178 full path</A
179></DT
180><DT
181><A
182HREF="#AEN121"
183>Use spopen() if external commands must be executed</A
184></DT
185><DT
186><A
187HREF="#AEN125"
188>Don't make temp files unless absolutely required</A
189></DT
190><DT
191><A
192HREF="#AEN128"
193>Don't be tricked into following symlinks</A
194></DT
195><DT
196><A
197HREF="#AEN131"
198>Validate all input</A
199></DT
200></DL
201></DD
202><DT
203><A
204HREF="#PERLPLUGIN"
205>Perl Plugins</A
206></DT
207><DT
208><A
209HREF="#RUNTIME"
210>Runtime Timeouts</A
211></DT
212><DD
213><DL
214><DT
215><A
216HREF="#AEN165"
217>Use DEFAULT_SOCKET_TIMEOUT</A
218></DT
219><DT
220><A
221HREF="#AEN168"
222>Add alarms to network plugins</A
223></DT
224></DL
225></DD
226><DT
227><A
228HREF="#PLUGOPTIONS"
229>Plugin Options</A
230></DT
231><DD
232><DL
233><DT
234><A
235HREF="#AEN174"
236>Option Processing</A
237></DT
238><DT
239><A
240HREF="#AEN187"
241>Plugins with more than one type of threshold, or with
242 threshold ranges</A
243></DT
244></DL
245></DD
246><DT
247><A
248HREF="#SUBMITTINGCHANGES"
249>New submissions and patches</A
250></DT
251></DL
252></DD
253></DL
254></DIV
255><DIV
256CLASS="PREFACE"
257><HR><H1
258><A
259NAME="PREFACE"
260>About the guidelines</A
261></H1
262><P
263>The purpose of this guidelines is to provide a reference for
264 the plug-in developers and encourage the standarization of the
265 different kind of plug-ins: C, shell, perl, python, etc.</P
266><DIV
267CLASS="SECTION"
268><HR><H1
269CLASS="SECTION"
270><A
271NAME="AEN51"
272>Copyright</A
273></H1
274><P
275>Nagios Plug-in Development Guidelines Copyright (C) 2000 2001
276 2002
277 Karl DeBisschop, Ethan Galstad, Hugo Gayosso, Stanley Hopcroft,
278 Subhendu Ghosh</P
279><P
280>Permission is granted to make and distribute verbatim
281 copies of this manual provided the copyright notice and this
282 permission notice are preserved on all copies.</P
283><P
284>The plugins themselves are copyrighted by their respective
285 authors.</P
286></DIV
287></DIV
288><DIV
289CLASS="ARTICLE"
290><DIV
291CLASS="TOC"
292><DL
293><DT
294><B
295>Table of Contents</B
296></DT
297><DT
298><A
299HREF="#PLUGOUTPUT"
300>Plugin Output for Nagios</A
301></DT
302><DT
303><A
304HREF="#SYSCMDAUXFILES"
305>System Commands and Auxiliary Files</A
306></DT
307><DT
308><A
309HREF="#PERLPLUGIN"
310>Perl Plugins</A
311></DT
312><DT
313><A
314HREF="#RUNTIME"
315>Runtime Timeouts</A
316></DT
317><DT
318><A
319HREF="#PLUGOPTIONS"
320>Plugin Options</A
321></DT
322><DT
323><A
324HREF="#SUBMITTINGCHANGES"
325>New submissions and patches</A
326></DT
327></DL
328></DIV
329><DIV
330CLASS="SECTION"
331><H1
332CLASS="SECTION"
333><A
334NAME="PLUGOUTPUT"
335>Plugin Output for Nagios</A
336></H1
337><P
338>You should always print something to STDOUT that tells if the
339 service is working or why its failing. Try to keep the output short -
340 probably less that 80 characters. Remember that you ideally would like
341 the entire output to appear in a pager message, which will get chopped
342 off after a certain length.</P
343><DIV
344CLASS="SECTION"
345><HR><H2
346CLASS="SECTION"
347><A
348NAME="AEN60"
349>Print only one line of text</A
350></H2
351><P
352>Nagios will only grab the first line of text from STDOUT
353 when it notifies contacts about potential problems. If you print
354 multiple lines, you're out of luck. Remember, keep it short and
355 to the point.</P
356></DIV
357><DIV
358CLASS="SECTION"
359><HR><H2
360CLASS="SECTION"
361><A
362NAME="AEN63"
363>Screen Output</A
364></H2
365><P
366>The plug-in should print the diagnostic and just the
367 synopsis part of the help message. A well written plugin would
368 then have --help as a way to get the verbose help.</P
369><P
370>Code and output should try to respect the 80x25 size of a
371 crt (remember when fixing stuff in the server room!)</P
372></DIV
373><DIV
374CLASS="SECTION"
375><HR><H2
376CLASS="SECTION"
377><A
378NAME="AEN67"
379>Return the proper status code</A
380></H2
381><P
382>See <A
383HREF="#RETURNCODES"
384>Table 1 in the section called <I
385>Plugin Return Codes</I
386></A
387> below
388 for the numeric values of status codes and their
389 description. Remember to return an UNKNOWN state if bogus or
390 invalid command line arguments are supplied or it you are unable
391 to check the service.</P
392></DIV
393><DIV
394CLASS="SECTION"
395><HR><H2
396CLASS="SECTION"
397><A
398NAME="AEN71"
399>Plugin Return Codes</A
400></H2
401><P
402>The return codes below are based on the POSIX spec of returning
403 a positive value. Netsaint prior to v0.0.7 supported non-POSIX
404 compliant return code of "-1" for unknown. Nagios supports POSIX return
405 codes by default.</P
406><P
407>Note: Some plugins will on occasion print on STDOUT that an error
408 occurred and error code is 138 or 255 or some such number. These
409 are usually caused by plugins using system commands and having not
410 enough checks to catch unexpected output. Developers should include a
411 default catch-all for system command output that returns an UNKOWN
412 return code.</P
413><DIV
414CLASS="TABLE"
415><A
416NAME="RETURNCODES"
417></A
418><P
419><B
420>Table 1. Plugin Return Codes</B
421></P
422><TABLE
423BORDER="1"
424BGCOLOR="#E0E0E0"
425CELLSPACING="0"
426CELLPADDING="4"
427CLASS="CALSTABLE"
428><THEAD
429><TR
430><TH
431ALIGN="LEFT"
432VALIGN="TOP"
433><P
434>Numeric Value</P
435></TH
436><TH
437ALIGN="LEFT"
438VALIGN="TOP"
439><P
440>Service Status</P
441></TH
442><TH
443ALIGN="LEFT"
444VALIGN="TOP"
445><P
446>Status Description</P
447></TH
448></TR
449></THEAD
450><TBODY
451><TR
452><TD
453ALIGN="CENTER"
454VALIGN="TOP"
455><P
456>0</P
457></TD
458><TD
459ALIGN="LEFT"
460VALIGN="MIDDLE"
461><P
462>OK</P
463></TD
464><TD
465ALIGN="LEFT"
466VALIGN="TOP"
467><P
468>The plugin was able to check the service and it
469 appeared to be functioning properly</P
470></TD
471></TR
472><TR
473><TD
474ALIGN="CENTER"
475VALIGN="TOP"
476><P
477>1</P
478></TD
479><TD
480ALIGN="LEFT"
481VALIGN="MIDDLE"
482><P
483>Warning</P
484></TD
485><TD
486ALIGN="LEFT"
487VALIGN="TOP"
488><P
489>The plugin was able to check the service, but it
490 appeared to be above some "warning" threshold or did not appear
491 to be working properly</P
492></TD
493></TR
494><TR
495><TD
496ALIGN="CENTER"
497VALIGN="TOP"
498><P
499>2</P
500></TD
501><TD
502ALIGN="LEFT"
503VALIGN="MIDDLE"
504><P
505>Critical</P
506></TD
507><TD
508ALIGN="LEFT"
509VALIGN="TOP"
510><P
511>The plugin detected that either the service was not
512 running or it was above some "critical" threshold</P
513></TD
514></TR
515><TR
516><TD
517ALIGN="CENTER"
518VALIGN="TOP"
519><P
520>3</P
521></TD
522><TD
523ALIGN="LEFT"
524VALIGN="MIDDLE"
525><P
526>Unknown</P
527></TD
528><TD
529ALIGN="LEFT"
530VALIGN="TOP"
531><P
532>Invalid command line arguments were supplied to the
533 plugin or the plugin was unable to check the status of the given
534 hosts/service</P
535></TD
536></TR
537></TBODY
538></TABLE
539></DIV
540></DIV
541></DIV
542><DIV
543CLASS="SECTION"
544><HR><H1
545CLASS="SECTION"
546><A
547NAME="SYSCMDAUXFILES"
548>System Commands and Auxiliary Files</A
549></H1
550><DIV
551CLASS="SECTION"
552><H2
553CLASS="SECTION"
554><A
555NAME="AEN117"
556>Don't execute system commands without specifying their
557 full path</A
558></H2
559><P
560>Don't use exec(), popen(), etc. to execute external
561 commands without explicity using the full path of the external
562 program.</P
563><P
564>Doing otherwise makes the plugin vulnerable to hijacking
565 by a trojan horse earlier in the search path. See the main
566 plugin distribution for examples on how this is done.</P
567></DIV
568><DIV
569CLASS="SECTION"
570><HR><H2
571CLASS="SECTION"
572><A
573NAME="AEN121"
574>Use spopen() if external commands must be executed</A
575></H2
576><P
577>If you have to execute external commands from within your
578 plugin and you're writing it in C, use the spopen() function
579 that Karl DeBisschop has written.</P
580><P
581>The code for spopen() and spclose() is included with the
582 core plugin distribution.</P
583></DIV
584><DIV
585CLASS="SECTION"
586><HR><H2
587CLASS="SECTION"
588><A
589NAME="AEN125"
590>Don't make temp files unless absolutely required</A
591></H2
592><P
593>If temp files are needed, make sure that the plugin will
594 fail cleanly if the file can't be written (e.g., too few file
595 handles, out of disk space, incorrect permissions, etc.) and
596 delete the temp file when processing is complete.</P
597></DIV
598><DIV
599CLASS="SECTION"
600><HR><H2
601CLASS="SECTION"
602><A
603NAME="AEN128"
604>Don't be tricked into following symlinks</A
605></H2
606><P
607>If your plugin opens any files, take steps to ensure that
608 you are not following a symlink to another location on the
609 system.</P
610></DIV
611><DIV
612CLASS="SECTION"
613><HR><H2
614CLASS="SECTION"
615><A
616NAME="AEN131"
617>Validate all input</A
618></H2
619><P
620>use routines in utils.c or utils.pm and write more as needed</P
621></DIV
622></DIV
623><DIV
624CLASS="SECTION"
625><HR><H1
626CLASS="SECTION"
627><A
628NAME="PERLPLUGIN"
629>Perl Plugins</A
630></H1
631><P
632>Perl plugins are coded a little more defensively than other
633 plugins because of embedded Perl. When configured as such, embedded
634 Perl Nagios (ePN) requires stricter use of the some of Perl's features.
635 This section outlines some of the steps needed to use ePN
636 effectively.</P
637><P
638></P
639><OL
640TYPE="1"
641><LI
642><P
643> Do not use BEGIN and END blocks since they will be called
644 the first time and when Nagios shuts down with Embedded Perl (ePN). In
645 particular, do not use BEGIN blocks to initialize variables.</P
646></LI
647><LI
648><P
649>To use utils.pm, you need to provide a full path to the
650 module in order for it to work with ePN.</P
651><P
652CLASS="LITERALLAYOUT"
653> &nbsp;&nbsp;e.g.<br>
654 use&nbsp;lib&nbsp;"/usr/local/nagios/libexec";<br>
655 use&nbsp;utils&nbsp;qw(...);<br>
656 &nbsp;&nbsp;</P
657></LI
658><LI
659><P
660>Perl scripts should be called with "-w"</P
661></LI
662><LI
663><P
664>All Perl plugins must compile cleanly under "use strict" - i.e. at
665 least explicitly package names as in "$main::x" or predeclare every
666 variable. </P
667><P
668>Explicitly initialize each varialable in use. Otherwise with
669 caching enabled, the plugin will not be recompilied each time, and
670 therefore Perl will not reinitialize all the variables. All old
671 variable values will still be in effect.</P
672></LI
673><LI
674><P
675>Do not use &#60; DATA &#62; (these simply do not compile under ePN).</P
676></LI
677><LI
678><P
679>Do not use named subroutines</P
680></LI
681><LI
682><P
683>If writing to a file (perhaps recording
684 performance data) explicitly close close it. The plugin never
685 calls <I
686CLASS="EMPHASIS"
687>exit</I
688>; that is caught by
689 p1.pl, so output streams are never closed.</P
690></LI
691><LI
692><P
693>As in <A
694HREF="#RUNTIME"
695>the section called <I
696>Runtime Timeouts</I
697></A
698> all plugins need
699 to monitor their runtime, specially if they are using network
700 resources. Use of the <I
701CLASS="EMPHASIS"
702>alarm</I
703> is recommended.
704 Plugins may import a default time out ($TIMEOUT) from utils.pm.
705 </P
706></LI
707><LI
708><P
709>Perl plugins should import %ERRORS from utils.pm
710 and then "exit $ERRORS{'OK'}" rather than "exit 0"
711 </P
712></LI
713></OL
714></DIV
715><DIV
716CLASS="SECTION"
717><HR><H1
718CLASS="SECTION"
719><A
720NAME="RUNTIME"
721>Runtime Timeouts</A
722></H1
723><P
724>Plugins have a very limited runtime - typically 10 sec.
725 As a result, it is very important for plugins to maintain internal
726 code to exit if runtime exceeds a threshold. </P
727><P
728>All plugins should timeout gracefully, not just networking
729 plugins. For instance, df may lock if you have automounted
730 drives and your network fails - but on first glance, who'd think
731 df could lock up like that. Plus, it should just be more error
732 resistant to be able to time out rather than consume
733 resources.</P
734><DIV
735CLASS="SECTION"
736><HR><H2
737CLASS="SECTION"
738><A
739NAME="AEN165"
740>Use DEFAULT_SOCKET_TIMEOUT</A
741></H2
742><P
743>All network plugins should use DEFAULT_SOCKET_TIMEOUT to timeout</P
744></DIV
745><DIV
746CLASS="SECTION"
747><HR><H2
748CLASS="SECTION"
749><A
750NAME="AEN168"
751>Add alarms to network plugins</A
752></H2
753><P
754>If you write a plugin which communicates with another
755 networked host, you should make sure to set an alarm() in your
756 code that prevents the plugin from hanging due to abnormal
757 socket closures, etc. Nagios takes steps to protect itself
758 against unruly plugins that timeout, but any plugins you create
759 should be well behaved on their own.</P
760></DIV
761></DIV
762><DIV
763CLASS="SECTION"
764><HR><H1
765CLASS="SECTION"
766><A
767NAME="PLUGOPTIONS"
768>Plugin Options</A
769></H1
770><P
771>A well written plugin should have --help as a way to get
772 verbose help. Code and output should try to respect the 80x25 size of a
773 crt (remember when fixing stuff in the server room!)</P
774><DIV
775CLASS="SECTION"
776><HR><H2
777CLASS="SECTION"
778><A
779NAME="AEN174"
780>Option Processing</A
781></H2
782><P
783>For plugins written in C, we recommend the C standard
784 getopt library for short options. If using getopt_long, check to
785 be sure that HAVE_GETOPT_H is defined (configure checks this and
786 sets the #define in common/config.h).</P
787><P
788>For plugins written in Perl, we recommend Getopt::Long module.</P
789><P
790>Positional arguments are strongly discouraged.</P
791><P
792>There are a few reserved options that should not be used
793 for other purposes:</P
794><P
795CLASS="LITERALLAYOUT"
796>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-V&nbsp;version&nbsp;(--version)<br>
797&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-h&nbsp;help&nbsp;(--help)<br>
798&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-t&nbsp;timeout&nbsp;(--timeout)<br>
799&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-w&nbsp;warning&nbsp;threshold&nbsp;(--warning)<br>
800&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-c&nbsp;critical&nbsp;threshold&nbsp;(--critical)<br>
801&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-H&nbsp;hostname&nbsp;(--hostname)<br>
802 </P
803><P
804>In addition to the reserved options above, some other standard options are:</P
805><P
806CLASS="LITERALLAYOUT"
807>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-C&nbsp;SNMP&nbsp;community&nbsp;(--community)<br>
808&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-a&nbsp;authentication&nbsp;password&nbsp;(--authentication)<br>
809&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-l&nbsp;login&nbsp;name&nbsp;(--logname)<br>
810&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-p&nbsp;port&nbsp;or&nbsp;password&nbsp;(--port&nbsp;or&nbsp;--passwd/--password)monitors&nbsp;operational<br>
811&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-u&nbsp;url&nbsp;or&nbsp;username&nbsp;(--url&nbsp;or&nbsp;--username)<br>
812 </P
813><P
814>Look at check_pgsql and check_procs to see how I currently
815 think this can work. Standard options are:</P
816><P
817>The option -V or --version should be present in all
818 plugins. For C plugins it should result in a call to print_revision, a
819 function in utils.c which takes two character arguments, the
820 command name and the plugin revision.</P
821><P
822>The -? option, or any other unparsable set of options,
823 should print out a short usage statement. Character width should
824 be 80 and less and no more that 23 lines should be printed (it
825 should display cleanly on a dumb terminal in a server
826 room).</P
827><P
828>The option -h or --help should be present in all plugins.
829 In C plugins, it should result in a call to print_help (or
830 equivalent). The function print_help should call print_revision,
831 then print_usage, then should provide detailed
832 help. Help text should fit on an 80-character width display, but
833 may run as many lines as needed.</P
834></DIV
835><DIV
836CLASS="SECTION"
837><HR><H2
838CLASS="SECTION"
839><A
840NAME="AEN187"
841>Plugins with more than one type of threshold, or with
842 threshold ranges</A
843></H2
844><P
845>Old style was to do things like -ct for critical time and
846 -cv for critical value. That goes out the window with POSIX
847 getopt. The allowable alternatves are:</P
848><P
849></P
850><OL
851TYPE="1"
852><LI
853><P
854>long options like -critical-time (or -ct and -cv, I
855 suppose).</P
856></LI
857><LI
858><P
859>repeated options like `check_load -w 10 -w 6 -w 4 -c
860 16 -c 10 -c 10`</P
861></LI
862><LI
863><P
864>for brevity, the above can be expressed as `check_load
865 -w 10,6,4 -c 16,10,10`</P
866></LI
867><LI
868><P
869>ranges are expressed with colons as in `check_procs -C
870 httpd -w 1:20 -c 1:30` which will warn above 20 instances,
871 and critical at 0 and above 30</P
872></LI
873><LI
874><P
875>lists are expressed with commas, so Jacob's check_nmap
876 uses constructs like '-p 1000,1010,1050:1060,2000'</P
877></LI
878><LI
879><P
880>If possible when writing lists, use tokens to make the
881 list easy to remember and non-order dependent - so
882 check_disk uses '-c 10000,10%' so that it is clear which is
883 the precentage and which is the KB values (note that due to
884 my own lack of foresight, that used to be '-c 10000:10%' but
885 such constructs should all be changed for consistency,
886 though providing reverse compatibility is fairly
887 easy).</P
888></LI
889></OL
890><P
891>As always, comments are welcome - making this consistent
892 without a host of long options was quite a hassle, and I would
893 suspect that there are flaws in this strategy. Perhaps clear
894 long-options is the most important of the above choices, but not
895 all POSIX systems have C libraries for long options, so the
896 short forms must exist as well.</P
897></DIV
898></DIV
899><DIV
900CLASS="SECTION"
901><HR><H1
902CLASS="SECTION"
903><A
904NAME="SUBMITTINGCHANGES"
905>New submissions and patches</A
906></H1
907><P
908>If you would like other to use your plugins and have it included in
909 the standard distribution, please include patches for the relavant
910 configuration files, in particular "configure.in" Otherwise submitted
911 plugins will be included in the contrib directory.</P
912><P
913>Plugins in the contrib directory are going to be migrated to the
914 standard plugins/plugin-scripts directory as time permits and per user
915 requests</P
916><P
917>Patches should be submitted via the SourceForge and be announced to
918 the mailing list.</P
919><P
920>For new plugins, provide a diff to add to the EXTRAS list (configure.in)
921 unless you are fairly sure that the plugin will work for all platforms with
922 no non-standard software added.</P
923><P
924>If possible please submit a test harness. Documentation on sample
925 tests coming soon.</P
926></DIV
927></DIV
928></DIV
929></BODY
930></HTML
931> \ No newline at end of file
diff --git a/doc/developer-guidelines.sgml b/doc/developer-guidelines.sgml
new file mode 100644
index 00000000..42ad8964
--- /dev/null
+++ b/doc/developer-guidelines.sgml
@@ -0,0 +1,483 @@
1<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook V4.1//EN">
2<book>
3 <title>Nagios Plug-in Developer Guidelines</title>
4
5 <bookinfo>
6 <authorgroup>
7 <author>
8 <firstname>Karl</firstname>
9 <surname>DeBisschop</surname>
10 <affiliation>
11 <address><email>karl@debisschop.net</email></address>
12 </affiliation>
13 </author>
14
15 <author>
16 <firstname>Ethan</firstname>
17 <surname>Galstad</surname>
18 <authorblurb>
19 <para>Author of Nagios</para>
20 <para><ulink url="http://www.nagios.org"></ulink></para>
21 </authorblurb>
22 <affiliation>
23 <address><email>netsaint@linuxbox.com</email></address>
24 </affiliation>
25 </author>
26
27 <author>
28 <firstname>Hugo</firstname>
29 <surname>Gayosso</surname>
30 <affiliation>
31 <address><email>hgayosso@gnu.org</email></address>
32 </affiliation>
33 </author>
34
35
36 <author>
37 <firstname>Subhendu</firstname>
38 <surname>Ghosh</surname>
39 <affiliation>
40 <address><email>sghosh@sourceforge.net</email></address>
41 </affiliation>
42 </author>
43
44 <author>
45 <firstname>Stanley</firstname>
46 <surname>Hopcroft</surname>
47 <affiliation>
48 <address><email>stanleyhopcroft@sourceforge.net</email></address>
49 </affiliation>
50 </author>
51
52 </authorgroup>
53
54 <pubdate>2002</pubdate>
55 <title>Nagios plug-in development guidelines</title>
56
57 <revhistory>
58 <revision>
59 <revnumber>0.4</revnumber>
60 <date>2 May 2002</date>
61 </revision>
62 </revhistory>
63
64 <copyright>
65 <year>2000 2001 2002</year>
66 <holder>Karl DeBisschop, Ethan Galstad,
67 Hugo Gayosso, Stanley Hopcroft, Subhendu Ghosh</holder>
68 </copyright>
69
70</bookinfo>
71
72
73 <preface id=preface>
74 <title>About the guidelines</title>
75
76 <para>The purpose of this guidelines is to provide a reference for
77 the plug-in developers and encourage the standarization of the
78 different kind of plug-ins: C, shell, perl, python, etc.</para>
79
80
81 <section> <title>Copyright</title>
82
83 <para>Nagios Plug-in Development Guidelines Copyright (C) 2000 2001
84 2002
85 Karl DeBisschop, Ethan Galstad, Hugo Gayosso, Stanley Hopcroft,
86 Subhendu Ghosh</para>
87
88 <para>Permission is granted to make and distribute verbatim
89 copies of this manual provided the copyright notice and this
90 permission notice are preserved on all copies.</para>
91
92 <para>The plugins themselves are copyrighted by their respective
93 authors.</para>
94
95 </section>
96</preface>
97
98<article>
99<section id="PlugOutput"><title>Plugin Output for Nagios</title>
100
101 <para>You should always print something to STDOUT that tells if the
102 service is working or why its failing. Try to keep the output short -
103 probably less that 80 characters. Remember that you ideally would like
104 the entire output to appear in a pager message, which will get chopped
105 off after a certain length.</para>
106
107 <section><title>Print only one line of text</title>
108 <para>Nagios will only grab the first line of text from STDOUT
109 when it notifies contacts about potential problems. If you print
110 multiple lines, you're out of luck. Remember, keep it short and
111 to the point.</para>
112 </section>
113
114 <section><title>Screen Output</title>
115 <para>The plug-in should print the diagnostic and just the
116 synopsis part of the help message. A well written plugin would
117 then have --help as a way to get the verbose help.</para>
118 <para>Code and output should try to respect the 80x25 size of a
119 crt (remember when fixing stuff in the server room!)</para>
120 </section>
121
122 <section><title>Return the proper status code</title>
123 <para>See <xref linkend="ReturnCodes"> below
124 for the numeric values of status codes and their
125 description. Remember to return an UNKNOWN state if bogus or
126 invalid command line arguments are supplied or it you are unable
127 to check the service.</para>
128 </section>
129
130 <section><title>Plugin Return Codes</title>
131 <para>The return codes below are based on the POSIX spec of returning
132 a positive value. Netsaint prior to v0.0.7 supported non-POSIX
133 compliant return code of "-1" for unknown. Nagios supports POSIX return
134 codes by default.</para>
135
136 <para>Note: Some plugins will on occasion print on STDOUT that an error
137 occurred and error code is 138 or 255 or some such number. These
138 are usually caused by plugins using system commands and having not
139 enough checks to catch unexpected output. Developers should include a
140 default catch-all for system command output that returns an UNKOWN
141 return code.</para>
142
143 <table id="ReturnCodes"><title>Plugin Return Codes</title>
144 <tgroup cols="3">
145 <thead>
146 <row>
147 <entry><para>Numeric Value</para></entry>
148 <entry><para>Service Status</para></entry>
149 <entry><para>Status Description</para></entry>
150 </row>
151 </thead>
152 <tbody>
153 <row>
154 <entry align=center><para>0</para></entry>
155 <entry valign=middle><para>OK</para></entry>
156 <entry><para>The plugin was able to check the service and it
157 appeared to be functioning properly</para></entry>
158 </row>
159 <row>
160 <entry align=center><para>1</para></entry>
161 <entry valign=middle><para>Warning</para></entry>
162 <entry><para>The plugin was able to check the service, but it
163 appeared to be above some "warning" threshold or did not appear
164 to be working properly</para></entry>
165 </row>
166 <row>
167 <entry align=center><para>2</para></entry>
168 <entry valign=middle><para>Critical</para></entry>
169 <entry><para>The plugin detected that either the service was not
170 running or it was above some "critical" threshold</para></entry>
171 </row>
172 <row>
173 <entry align=center><para>3</para></entry>
174 <entry valign=middle><para>Unknown</para></entry>
175 <entry><para>Invalid command line arguments were supplied to the
176 plugin or the plugin was unable to check the status of the given
177 hosts/service</para></entry>
178 </row>
179 </tbody>
180 </tgroup>
181 </table>
182
183
184 </section>
185
186
187</section>
188
189<section id="SysCmdAuxFiles"><title>System Commands and Auxiliary Files</title>
190
191 <section><title>Don't execute system commands without specifying their
192 full path</title>
193 <para>Don't use exec(), popen(), etc. to execute external
194 commands without explicity using the full path of the external
195 program.</para>
196
197 <para>Doing otherwise makes the plugin vulnerable to hijacking
198 by a trojan horse earlier in the search path. See the main
199 plugin distribution for examples on how this is done.</para>
200 </section>
201
202 <section><title>Use spopen() if external commands must be executed</title>
203
204 <para>If you have to execute external commands from within your
205 plugin and you're writing it in C, use the spopen() function
206 that Karl DeBisschop has written.</para>
207
208 <para>The code for spopen() and spclose() is included with the
209 core plugin distribution.</para>
210 </section>
211
212 <section><title>Don't make temp files unless absolutely required</title>
213
214 <para>If temp files are needed, make sure that the plugin will
215 fail cleanly if the file can't be written (e.g., too few file
216 handles, out of disk space, incorrect permissions, etc.) and
217 delete the temp file when processing is complete.</para>
218 </section>
219
220 <section><title>Don't be tricked into following symlinks</title>
221
222 <para>If your plugin opens any files, take steps to ensure that
223 you are not following a symlink to another location on the
224 system.</para>
225 </section>
226
227 <section><title>Validate all input</title>
228
229 <para>use routines in utils.c or utils.pm and write more as needed</para>
230 </section>
231
232</section>
233
234
235
236
237<section id="PerlPlugin"><title>Perl Plugins</title>
238
239 <para>Perl plugins are coded a little more defensively than other
240 plugins because of embedded Perl. When configured as such, embedded
241 Perl Nagios (ePN) requires stricter use of the some of Perl's features.
242 This section outlines some of the steps needed to use ePN
243 effectively.</para>
244
245 <orderedlist>
246
247 <listitem><para> Do not use BEGIN and END blocks since they will be called
248 the first time and when Nagios shuts down with Embedded Perl (ePN). In
249 particular, do not use BEGIN blocks to initialize variables.</para>
250 </listitem>
251
252 <listitem><para>To use utils.pm, you need to provide a full path to the
253 module in order for it to work with ePN.</para>
254
255 <literallayout>
256 e.g.
257 use lib "/usr/local/nagios/libexec";
258 use utils qw(...);
259 </literallayout>
260 </listitem>
261
262 <listitem><para>Perl scripts should be called with "-w"</para>
263 </listitem>
264
265 <listitem><para>All Perl plugins must compile cleanly under "use strict" - i.e. at
266 least explicitly package names as in "$main::x" or predeclare every
267 variable. </para>
268
269
270 <para>Explicitly initialize each varialable in use. Otherwise with
271 caching enabled, the plugin will not be recompilied each time, and
272 therefore Perl will not reinitialize all the variables. All old
273 variable values will still be in effect.</para>
274 </listitem>
275
276 <listitem><para>Do not use < DATA > (these simply do not compile under ePN).</para>
277 </listitem>
278
279 <listitem><para>Do not use named subroutines</para>
280 </listitem>
281
282 <listitem><para>If writing to a file (perhaps recording
283 performance data) explicitly close close it. The plugin never
284 calls <emphasis role=strong>exit</emphasis>; that is caught by
285 p1.pl, so output streams are never closed.</para>
286 </listitem>
287
288 <listitem><para>As in <xref linkend="runtime"> all plugins need
289 to monitor their runtime, specially if they are using network
290 resources. Use of the <emphasis>alarm</emphasis> is recommended.
291 Plugins may import a default time out ($TIMEOUT) from utils.pm.
292 </para>
293 </listitem>
294
295 <listitem><para>Perl plugins should import %ERRORS from utils.pm
296 and then "exit $ERRORS{'OK'}" rather than "exit 0"
297 </para>
298 </listitem>
299
300 </orderedlist>
301
302</section>
303
304<section id="runtime"><title>Runtime Timeouts</title>
305
306 <para>Plugins have a very limited runtime - typically 10 sec.
307 As a result, it is very important for plugins to maintain internal
308 code to exit if runtime exceeds a threshold. </para>
309
310 <para>All plugins should timeout gracefully, not just networking
311 plugins. For instance, df may lock if you have automounted
312 drives and your network fails - but on first glance, who'd think
313 df could lock up like that. Plus, it should just be more error
314 resistant to be able to time out rather than consume
315 resources.</para>
316
317 <section><title>Use DEFAULT_SOCKET_TIMEOUT</title>
318
319 <para>All network plugins should use DEFAULT_SOCKET_TIMEOUT to timeout</para>
320
321 </section>
322
323
324 <section><title>Add alarms to network plugins</title>
325
326 <para>If you write a plugin which communicates with another
327 networked host, you should make sure to set an alarm() in your
328 code that prevents the plugin from hanging due to abnormal
329 socket closures, etc. Nagios takes steps to protect itself
330 against unruly plugins that timeout, but any plugins you create
331 should be well behaved on their own.</para>
332
333 </section>
334
335
336
337</section>
338
339<section id="PlugOptions"><title>Plugin Options</title>
340
341 <para>A well written plugin should have --help as a way to get
342 verbose help. Code and output should try to respect the 80x25 size of a
343 crt (remember when fixing stuff in the server room!)</para>
344
345 <section><title>Option Processing</title>
346
347 <para>For plugins written in C, we recommend the C standard
348 getopt library for short options. If using getopt_long, check to
349 be sure that HAVE_GETOPT_H is defined (configure checks this and
350 sets the #define in common/config.h).</para>
351
352 <para>For plugins written in Perl, we recommend Getopt::Long module.</para>
353
354 <para>Positional arguments are strongly discouraged.</para>
355
356 <para>There are a few reserved options that should not be used
357 for other purposes:</para>
358
359 <literallayout>
360 -V version (--version)
361 -h help (--help)
362 -t timeout (--timeout)
363 -w warning threshold (--warning)
364 -c critical threshold (--critical)
365 -H hostname (--hostname)
366 </literallayout>
367
368 <para>In addition to the reserved options above, some other standard options are:</para>
369
370 <literallayout>
371 -C SNMP community (--community)
372 -a authentication password (--authentication)
373 -l login name (--logname)
374 -p port or password (--port or --passwd/--password)monitors operational
375 -u url or username (--url or --username)
376 </literallayout>
377
378 <para>Look at check_pgsql and check_procs to see how I currently
379 think this can work. Standard options are:</para>
380
381
382 <para>The option -V or --version should be present in all
383 plugins. For C plugins it should result in a call to print_revision, a
384 function in utils.c which takes two character arguments, the
385 command name and the plugin revision.</para>
386
387 <para>The -? option, or any other unparsable set of options,
388 should print out a short usage statement. Character width should
389 be 80 and less and no more that 23 lines should be printed (it
390 should display cleanly on a dumb terminal in a server
391 room).</para>
392
393 <para>The option -h or --help should be present in all plugins.
394 In C plugins, it should result in a call to print_help (or
395 equivalent). The function print_help should call print_revision,
396 then print_usage, then should provide detailed
397 help. Help text should fit on an 80-character width display, but
398 may run as many lines as needed.</para>
399
400 </section>
401
402 <section>
403 <title>Plugins with more than one type of threshold, or with
404 threshold ranges</title>
405
406 <para>Old style was to do things like -ct for critical time and
407 -cv for critical value. That goes out the window with POSIX
408 getopt. The allowable alternatves are:</para>
409
410 <orderedlist>
411 <listitem>
412 <para>long options like -critical-time (or -ct and -cv, I
413 suppose).</para>
414 </listitem>
415
416 <listitem>
417 <para>repeated options like `check_load -w 10 -w 6 -w 4 -c
418 16 -c 10 -c 10`</para>
419 </listitem>
420
421 <listitem>
422 <para>for brevity, the above can be expressed as `check_load
423 -w 10,6,4 -c 16,10,10`</para>
424 </listitem>
425
426 <listitem>
427 <para>ranges are expressed with colons as in `check_procs -C
428 httpd -w 1:20 -c 1:30` which will warn above 20 instances,
429 and critical at 0 and above 30</para>
430 </listitem>
431
432 <listitem>
433 <para>lists are expressed with commas, so Jacob's check_nmap
434 uses constructs like '-p 1000,1010,1050:1060,2000'</para>
435 </listitem>
436
437 <listitem>
438 <para>If possible when writing lists, use tokens to make the
439 list easy to remember and non-order dependent - so
440 check_disk uses '-c 10000,10%' so that it is clear which is
441 the precentage and which is the KB values (note that due to
442 my own lack of foresight, that used to be '-c 10000:10%' but
443 such constructs should all be changed for consistency,
444 though providing reverse compatibility is fairly
445 easy).</para>
446 </listitem>
447
448 </orderedlist>
449
450 <para>As always, comments are welcome - making this consistent
451 without a host of long options was quite a hassle, and I would
452 suspect that there are flaws in this strategy. Perhaps clear
453 long-options is the most important of the above choices, but not
454 all POSIX systems have C libraries for long options, so the
455 short forms must exist as well.</para>
456 </section>
457</section>
458
459<section id="SubmittingChanges"><title>New submissions and patches</title>
460
461 <para>If you would like other to use your plugins and have it included in
462 the standard distribution, please include patches for the relavant
463 configuration files, in particular "configure.in" Otherwise submitted
464 plugins will be included in the contrib directory.</para>
465
466 <para>Plugins in the contrib directory are going to be migrated to the
467 standard plugins/plugin-scripts directory as time permits and per user
468 requests</para>
469
470 <para>Patches should be submitted via the SourceForge and be announced to
471 the mailing list.</para>
472
473 <para>For new plugins, provide a diff to add to the EXTRAS list (configure.in)
474 unless you are fairly sure that the plugin will work for all platforms with
475 no non-standard software added.</para>
476
477 <para>If possible please submit a test harness. Documentation on sample
478 tests coming soon.</para>
479
480</section>
481</article>
482
483</book>