unbound_munin_ revision 356345
1#!/bin/sh 2# 3# plugin for munin to monitor usage of unbound servers. 4# To install copy this to /usr/local/share/munin/plugins/unbound_munin_ 5# and use munin-node-configure (--suggest, --shell). 6# 7# (C) 2008 W.C.A. Wijngaards. BSD Licensed. 8# 9# To install; enable statistics and unbound-control in unbound.conf 10# server: extended-statistics: yes 11# statistics-cumulative: no 12# statistics-interval: 0 13# remote-control: control-enable: yes 14# Run the command unbound-control-setup to generate the key files. 15# 16# Environment variables for this script 17# statefile - where to put temporary statefile. 18# unbound_conf - where the unbound.conf file is located. 19# unbound_control - where to find unbound-control executable. 20# spoof_warn - what level to warn about spoofing 21# spoof_crit - what level to crit about spoofing 22# 23# You can set them in your munin/plugin-conf.d/plugins.conf file 24# with: 25# [unbound*] 26# user root 27# env.statefile /usr/local/var/munin/plugin-state/unbound-state 28# env.unbound_conf /usr/local/etc/unbound/unbound.conf 29# env.unbound_control /usr/local/sbin/unbound-control 30# env.spoof_warn 1000 31# env.spoof_crit 100000 32# 33# This plugin can create different graphs depending on what name 34# you link it as (with ln -s) into the plugins directory 35# You can link it multiple times. 36# If you are only a casual user, the _hits and _by_type are most interesting, 37# possibly followed by _by_rcode. 38# 39# unbound_munin_hits - base volume, cache hits, unwanted traffic 40# unbound_munin_queue - to monitor the internal requestlist 41# unbound_munin_memory - memory usage 42# unbound_munin_by_type - incoming queries by type 43# unbound_munin_by_class - incoming queries by class 44# unbound_munin_by_opcode - incoming queries by opcode 45# unbound_munin_by_rcode - answers by rcode, validation status 46# unbound_munin_by_flags - incoming queries by flags 47# unbound_munin_histogram - histogram of query resolving times 48# 49# Magic markers - optional - used by installation scripts and 50# munin-config: (originally contrib family but munin-node-configure ignores it) 51# 52#%# family=auto 53#%# capabilities=autoconf suggest 54 55# POD documentation 56: <<=cut 57=head1 NAME 58 59unbound_munin_ - Munin plugin to monitor the Unbound DNS resolver. 60 61=head1 APPLICABLE SYSTEMS 62 63System with unbound daemon. 64 65=head1 CONFIGURATION 66 67 [unbound*] 68 user root 69 env.statefile /usr/local/var/munin/plugin-state/unbound-state 70 env.unbound_conf /usr/local/etc/unbound/unbound.conf 71 env.unbound_control /usr/local/sbin/unbound-control 72 env.spoof_warn 1000 73 env.spoof_crit 100000 74 75Use the .env settings to override the defaults. 76 77=head1 USAGE 78 79Can be used to present different graphs. Use ln -s for that name in 80the plugins directory to enable the graph. 81unbound_munin_hits - base volume, cache hits, unwanted traffic 82unbound_munin_queue - to monitor the internal requestlist 83unbound_munin_memory - memory usage 84unbound_munin_by_type - incoming queries by type 85unbound_munin_by_class - incoming queries by class 86unbound_munin_by_opcode - incoming queries by opcode 87unbound_munin_by_rcode - answers by rcode, validation status 88unbound_munin_by_flags - incoming queries by flags 89unbound_munin_histogram - histogram of query resolving times 90 91=head1 AUTHOR 92 93Copyright 2008 W.C.A. Wijngaards 94 95=head1 LICENSE 96 97BSD 98 99=cut 100 101state=${statefile:-/usr/local/var/munin/plugin-state/unbound-state} 102conf=${unbound_conf:-/usr/local/etc/unbound/unbound.conf} 103ctrl=${unbound_control:-/usr/local/sbin/unbound-control} 104warn=${spoof_warn:-1000} 105crit=${spoof_crit:-100000} 106lock=$state.lock 107 108# number of seconds between polling attempts. 109# makes the statefile hang around for at least this many seconds, 110# so that multiple links of this script can share the results. 111lee=55 112 113# to keep things within 19 characters 114ABBREV="-e s/total/t/ -e s/thread/t/ -e s/num/n/ -e s/query/q/ -e s/answer/a/ -e s/unwanted/u/ -e s/requestlist/ql/ -e s/type/t/ -e s/class/c/ -e s/opcode/o/ -e s/rcode/r/ -e s/edns/e/ -e s/mem/m/ -e s/cache/c/ -e s/mod/m/" 115 116# get value from $1 into return variable $value 117get_value ( ) { 118 value="`grep '^'$1'=' $state | sed -e 's/^.*=//'`" 119 if test "$value"x = ""x; then 120 value="0" 121 fi 122} 123 124# download the state from the unbound server. 125get_state ( ) { 126 # obtain lock for fetching the state 127 # because there is a race condition in fetching and writing to file 128 129 # see if the lock is stale, if so, take it 130 if test -f $lock ; then 131 pid="`cat $lock 2>&1`" 132 kill -0 "$pid" >/dev/null 2>&1 133 if test $? -ne 0 -a "$pid" != $$ ; then 134 echo $$ >$lock 135 fi 136 fi 137 138 i=0 139 while test ! -f $lock || test "`cat $lock 2>&1`" != $$; do 140 while test -f $lock; do 141 # wait 142 i=`expr $i + 1` 143 if test $i -gt 1000; then 144 sleep 1; 145 fi 146 if test $i -gt 1500; then 147 echo "error locking $lock" "=" `cat $lock` 148 rm -f $lock 149 exit 1 150 fi 151 done 152 # try to get it 153 if echo $$ >$lock ; then : ; else break; fi 154 done 155 # do not refetch if the file exists and only LEE seconds old 156 if test -f $state; then 157 now=`date +%s` 158 get_value "time.now" 159 value="`echo $value | sed -e 's/\..*$//'`" 160 if test $now -lt `expr $value + $lee`; then 161 rm -f $lock 162 return 163 fi 164 fi 165 $ctrl -c $conf stats > $state 166 if test $? -ne 0; then 167 echo "error retrieving data from unbound server" 168 rm -f $lock 169 exit 1 170 fi 171 rm -f $lock 172} 173 174if test "$1" = "autoconf" ; then 175 if test ! -f $conf; then 176 echo no "($conf does not exist)" 177 exit 1 178 fi 179 if test ! -d `dirname $state`; then 180 echo no "(`dirname $state` directory does not exist)" 181 exit 1 182 fi 183 echo yes 184 exit 0 185fi 186 187if test "$1" = "suggest" ; then 188 echo "hits" 189 echo "queue" 190 echo "memory" 191 echo "by_type" 192 echo "by_class" 193 echo "by_opcode" 194 echo "by_rcode" 195 echo "by_flags" 196 echo "histogram" 197 exit 0 198fi 199 200# determine my type, by name 201id=`echo $0 | sed -e 's/^.*unbound_munin_//'` 202if test "$id"x = ""x; then 203 # some default to keep people sane. 204 id="hits" 205fi 206 207# if $1 exists in statefile, config is echoed with label $2 208exist_config ( ) { 209 mn=`echo $1 | sed $ABBREV | tr . _` 210 if grep '^'$1'=' $state >/dev/null 2>&1; then 211 echo "$mn.label $2" 212 echo "$mn.min 0" 213 echo "$mn.type ABSOLUTE" 214 fi 215} 216 217# print label and min 0 for a name $1 in unbound format 218p_config ( ) { 219 mn=`echo $1 | sed $ABBREV | tr . _` 220 echo $mn.label "$2" 221 echo $mn.min 0 222 echo $mn.type $3 223} 224 225if test "$1" = "config" ; then 226 if test ! -f $state; then 227 get_state 228 fi 229 case $id in 230 hits) 231 echo "graph_title Unbound DNS traffic and cache hits" 232 echo "graph_args --base 1000 -l 0" 233 echo "graph_vlabel queries / \${graph_period}" 234 echo "graph_scale no" 235 echo "graph_category DNS" 236 for x in `grep "^thread[0-9][0-9]*\.num\.queries=" $state | 237 sed -e 's/=.*//'`; do 238 exist_config $x "queries handled by `basename $x .num.queries`" 239 done 240 p_config "total.num.queries" "total queries from clients" "ABSOLUTE" 241 p_config "total.num.cachehits" "cache hits" "ABSOLUTE" 242 p_config "total.num.prefetch" "cache prefetch" "ABSOLUTE" 243 p_config "num.query.tcp" "TCP queries" "ABSOLUTE" 244 p_config "num.query.tcpout" "TCP out queries" "ABSOLUTE" 245 p_config "num.query.ipv6" "IPv6 queries" "ABSOLUTE" 246 p_config "unwanted.queries" "queries that failed acl" "ABSOLUTE" 247 p_config "unwanted.replies" "unwanted or unsolicited replies" "ABSOLUTE" 248 echo "u_replies.warning $warn" 249 echo "u_replies.critical $crit" 250 echo "graph_info DNS queries to the recursive resolver. The unwanted replies could be innocent duplicate packets, late replies, or spoof threats." 251 ;; 252 queue) 253 echo "graph_title Unbound requestlist size" 254 echo "graph_args --base 1000 -l 0" 255 echo "graph_vlabel number of queries" 256 echo "graph_scale no" 257 echo "graph_category DNS" 258 p_config "total.requestlist.avg" "Average size of queue on insert" "GAUGE" 259 p_config "total.requestlist.max" "Max size of queue (in 5 min)" "GAUGE" 260 p_config "total.requestlist.overwritten" "Number of queries replaced by new ones" "GAUGE" 261 p_config "total.requestlist.exceeded" "Number of queries dropped due to lack of space" "GAUGE" 262 echo "graph_info The queries that did not hit the cache and need recursion service take up space in the requestlist. If there are too many queries, first queries get overwritten, and at last resort dropped." 263 ;; 264 memory) 265 echo "graph_title Unbound memory usage" 266 echo "graph_args --base 1024 -l 0" 267 echo "graph_vlabel memory used in bytes" 268 echo "graph_category DNS" 269 p_config "mem.cache.rrset" "RRset cache memory" "GAUGE" 270 p_config "mem.cache.message" "Message cache memory" "GAUGE" 271 p_config "mem.mod.iterator" "Iterator module memory" "GAUGE" 272 p_config "mem.mod.validator" "Validator module and key cache memory" "GAUGE" 273 p_config "msg.cache.count" "msg cache count" "GAUGE" 274 p_config "rrset.cache.count" "rrset cache count" "GAUGE" 275 p_config "infra.cache.count" "infra cache count" "GAUGE" 276 p_config "key.cache.count" "key cache count" "GAUGE" 277 echo "graph_info The memory used by unbound." 278 ;; 279 by_type) 280 echo "graph_title Unbound DNS queries by type" 281 echo "graph_args --base 1000 -l 0" 282 echo "graph_vlabel queries / \${graph_period}" 283 echo "graph_scale no" 284 echo "graph_category DNS" 285 for x in `grep "^num.query.type" $state`; do 286 nm=`echo $x | sed -e 's/=.*$//'` 287 tp=`echo $nm | sed -e s/num.query.type.//` 288 p_config "$nm" "$tp" "ABSOLUTE" 289 done 290 echo "graph_info queries by DNS RR type queried for" 291 ;; 292 by_class) 293 echo "graph_title Unbound DNS queries by class" 294 echo "graph_args --base 1000 -l 0" 295 echo "graph_vlabel queries / \${graph_period}" 296 echo "graph_scale no" 297 echo "graph_category DNS" 298 for x in `grep "^num.query.class" $state`; do 299 nm=`echo $x | sed -e 's/=.*$//'` 300 tp=`echo $nm | sed -e s/num.query.class.//` 301 p_config "$nm" "$tp" "ABSOLUTE" 302 done 303 echo "graph_info queries by DNS RR class queried for." 304 ;; 305 by_opcode) 306 echo "graph_title Unbound DNS queries by opcode" 307 echo "graph_args --base 1000 -l 0" 308 echo "graph_vlabel queries / \${graph_period}" 309 echo "graph_scale no" 310 echo "graph_category DNS" 311 for x in `grep "^num.query.opcode" $state`; do 312 nm=`echo $x | sed -e 's/=.*$//'` 313 tp=`echo $nm | sed -e s/num.query.opcode.//` 314 p_config "$nm" "$tp" "ABSOLUTE" 315 done 316 echo "graph_info queries by opcode in the query packet." 317 ;; 318 by_rcode) 319 echo "graph_title Unbound DNS answers by return code" 320 echo "graph_args --base 1000 -l 0" 321 echo "graph_vlabel answer packets / \${graph_period}" 322 echo "graph_scale no" 323 echo "graph_category DNS" 324 for x in `grep "^num.answer.rcode" $state`; do 325 nm=`echo $x | sed -e 's/=.*$//'` 326 tp=`echo $nm | sed -e s/num.answer.rcode.//` 327 p_config "$nm" "$tp" "ABSOLUTE" 328 done 329 p_config "num.answer.secure" "answer secure" "ABSOLUTE" 330 p_config "num.answer.bogus" "answer bogus" "ABSOLUTE" 331 p_config "num.rrset.bogus" "num rrsets marked bogus" "ABSOLUTE" 332 echo "graph_info answers sorted by return value. rrsets bogus is the number of rrsets marked bogus per \${graph_period} by the validator" 333 ;; 334 by_flags) 335 echo "graph_title Unbound DNS incoming queries by flags" 336 echo "graph_args --base 1000 -l 0" 337 echo "graph_vlabel queries / \${graph_period}" 338 echo "graph_scale no" 339 echo "graph_category DNS" 340 p_config "num.query.flags.QR" "QR (query reply) flag" "ABSOLUTE" 341 p_config "num.query.flags.AA" "AA (auth answer) flag" "ABSOLUTE" 342 p_config "num.query.flags.TC" "TC (truncated) flag" "ABSOLUTE" 343 p_config "num.query.flags.RD" "RD (recursion desired) flag" "ABSOLUTE" 344 p_config "num.query.flags.RA" "RA (rec avail) flag" "ABSOLUTE" 345 p_config "num.query.flags.Z" "Z (zero) flag" "ABSOLUTE" 346 p_config "num.query.flags.AD" "AD (auth data) flag" "ABSOLUTE" 347 p_config "num.query.flags.CD" "CD (check disabled) flag" "ABSOLUTE" 348 p_config "num.query.edns.present" "EDNS OPT present" "ABSOLUTE" 349 p_config "num.query.edns.DO" "DO (DNSSEC OK) flag" "ABSOLUTE" 350 echo "graph_info This graphs plots the flags inside incoming queries. For example, if QR, AA, TC, RA, Z flags are set, the query can be rejected. RD, AD, CD and DO are legitimately set by some software." 351 ;; 352 histogram) 353 echo "graph_title Unbound DNS histogram of reply time" 354 echo "graph_args --base 1000 -l 0" 355 echo "graph_vlabel queries / \${graph_period}" 356 echo "graph_scale no" 357 echo "graph_category DNS" 358 echo hcache.label "cache hits" 359 echo hcache.min 0 360 echo hcache.type ABSOLUTE 361 echo hcache.draw AREA 362 echo hcache.colour 999999 363 echo h64ms.label "0 msec - 66 msec" 364 echo h64ms.min 0 365 echo h64ms.type ABSOLUTE 366 echo h64ms.draw STACK 367 echo h64ms.colour 0000FF 368 echo h128ms.label "66 msec - 131 msec" 369 echo h128ms.min 0 370 echo h128ms.type ABSOLUTE 371 echo h128ms.colour 1F00DF 372 echo h128ms.draw STACK 373 echo h256ms.label "131 msec - 262 msec" 374 echo h256ms.min 0 375 echo h256ms.type ABSOLUTE 376 echo h256ms.draw STACK 377 echo h256ms.colour 3F00BF 378 echo h512ms.label "262 msec - 524 msec" 379 echo h512ms.min 0 380 echo h512ms.type ABSOLUTE 381 echo h512ms.draw STACK 382 echo h512ms.colour 5F009F 383 echo h1s.label "524 msec - 1 sec" 384 echo h1s.min 0 385 echo h1s.type ABSOLUTE 386 echo h1s.draw STACK 387 echo h1s.colour 7F007F 388 echo h2s.label "1 sec - 2 sec" 389 echo h2s.min 0 390 echo h2s.type ABSOLUTE 391 echo h2s.draw STACK 392 echo h2s.colour 9F005F 393 echo h4s.label "2 sec - 4 sec" 394 echo h4s.min 0 395 echo h4s.type ABSOLUTE 396 echo h4s.draw STACK 397 echo h4s.colour BF003F 398 echo h8s.label "4 sec - 8 sec" 399 echo h8s.min 0 400 echo h8s.type ABSOLUTE 401 echo h8s.draw STACK 402 echo h8s.colour DF001F 403 echo h16s.label "8 sec - ..." 404 echo h16s.min 0 405 echo h16s.type ABSOLUTE 406 echo h16s.draw STACK 407 echo h16s.colour FF0000 408 echo "graph_info Histogram of the reply times for queries." 409 ;; 410 esac 411 412 exit 0 413fi 414 415# do the stats itself 416get_state 417 418# get the time elapsed 419get_value "time.elapsed" 420if test $value = 0 || test $value = "0.000000"; then 421 echo "error: time elapsed 0 or could not retrieve data" 422 exit 1 423fi 424elapsed="$value" 425 426# print value for $1 427print_value ( ) { 428 mn=`echo $1 | sed $ABBREV | tr . _` 429 get_value $1 430 echo "$mn.value" $value 431} 432 433# print value if line already found in $2 434print_value_line ( ) { 435 mn=`echo $1 | sed $ABBREV | tr . _` 436 value="`echo $2 | sed -e 's/^.*=//'`" 437 echo "$mn.value" $value 438} 439 440 441case $id in 442hits) 443 for x in `grep "^thread[0-9][0-9]*\.num\.queries=" $state | 444 sed -e 's/=.*//'` total.num.queries \ 445 total.num.cachehits total.num.prefetch num.query.tcp \ 446 num.query.tcpout num.query.ipv6 unwanted.queries \ 447 unwanted.replies; do 448 if grep "^"$x"=" $state >/dev/null 2>&1; then 449 print_value $x 450 fi 451 done 452 ;; 453queue) 454 for x in total.requestlist.avg total.requestlist.max \ 455 total.requestlist.overwritten total.requestlist.exceeded; do 456 print_value $x 457 done 458 ;; 459memory) 460 for x in mem.cache.rrset mem.cache.message mem.mod.iterator \ 461 mem.mod.validator msg.cache.count rrset.cache.count \ 462 infra.cache.count key.cache.count; do 463 print_value $x 464 done 465 ;; 466by_type) 467 for x in `grep "^num.query.type" $state`; do 468 nm=`echo $x | sed -e 's/=.*$//'` 469 print_value_line $nm $x 470 done 471 ;; 472by_class) 473 for x in `grep "^num.query.class" $state`; do 474 nm=`echo $x | sed -e 's/=.*$//'` 475 print_value_line $nm $x 476 done 477 ;; 478by_opcode) 479 for x in `grep "^num.query.opcode" $state`; do 480 nm=`echo $x | sed -e 's/=.*$//'` 481 print_value_line $nm $x 482 done 483 ;; 484by_rcode) 485 for x in `grep "^num.answer.rcode" $state`; do 486 nm=`echo $x | sed -e 's/=.*$//'` 487 print_value_line $nm $x 488 done 489 print_value "num.answer.secure" 490 print_value "num.answer.bogus" 491 print_value "num.rrset.bogus" 492 ;; 493by_flags) 494 for x in num.query.flags.QR num.query.flags.AA num.query.flags.TC num.query.flags.RD num.query.flags.RA num.query.flags.Z num.query.flags.AD num.query.flags.CD num.query.edns.present num.query.edns.DO; do 495 print_value $x 496 done 497 ;; 498histogram) 499 get_value total.num.cachehits 500 echo hcache.value $value 501 r=0 502 for x in histogram.000000.000000.to.000000.000001 \ 503 histogram.000000.000001.to.000000.000002 \ 504 histogram.000000.000002.to.000000.000004 \ 505 histogram.000000.000004.to.000000.000008 \ 506 histogram.000000.000008.to.000000.000016 \ 507 histogram.000000.000016.to.000000.000032 \ 508 histogram.000000.000032.to.000000.000064 \ 509 histogram.000000.000064.to.000000.000128 \ 510 histogram.000000.000128.to.000000.000256 \ 511 histogram.000000.000256.to.000000.000512 \ 512 histogram.000000.000512.to.000000.001024 \ 513 histogram.000000.001024.to.000000.002048 \ 514 histogram.000000.002048.to.000000.004096 \ 515 histogram.000000.004096.to.000000.008192 \ 516 histogram.000000.008192.to.000000.016384 \ 517 histogram.000000.016384.to.000000.032768 \ 518 histogram.000000.032768.to.000000.065536; do 519 get_value $x 520 r=`expr $r + $value` 521 done 522 echo h64ms.value $r 523 get_value histogram.000000.065536.to.000000.131072 524 echo h128ms.value $value 525 get_value histogram.000000.131072.to.000000.262144 526 echo h256ms.value $value 527 get_value histogram.000000.262144.to.000000.524288 528 echo h512ms.value $value 529 get_value histogram.000000.524288.to.000001.000000 530 echo h1s.value $value 531 get_value histogram.000001.000000.to.000002.000000 532 echo h2s.value $value 533 get_value histogram.000002.000000.to.000004.000000 534 echo h4s.value $value 535 get_value histogram.000004.000000.to.000008.000000 536 echo h8s.value $value 537 r=0 538 for x in histogram.000008.000000.to.000016.000000 \ 539 histogram.000016.000000.to.000032.000000 \ 540 histogram.000032.000000.to.000064.000000 \ 541 histogram.000064.000000.to.000128.000000 \ 542 histogram.000128.000000.to.000256.000000 \ 543 histogram.000256.000000.to.000512.000000 \ 544 histogram.000512.000000.to.001024.000000 \ 545 histogram.001024.000000.to.002048.000000 \ 546 histogram.002048.000000.to.004096.000000 \ 547 histogram.004096.000000.to.008192.000000 \ 548 histogram.008192.000000.to.016384.000000 \ 549 histogram.016384.000000.to.032768.000000 \ 550 histogram.032768.000000.to.065536.000000 \ 551 histogram.065536.000000.to.131072.000000 \ 552 histogram.131072.000000.to.262144.000000 \ 553 histogram.262144.000000.to.524288.000000; do 554 get_value $x 555 r=`expr $r + $value` 556 done 557 echo h16s.value $r 558 ;; 559esac 560