1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
|
/*
* Copyright 2013-2016 Fabian Groffen
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <time.h>
#include <pthread.h>
#include <assert.h>
#include "relay.h"
#include "dispatcher.h"
#include "server.h"
#include "aggregator.h"
#include "collector.h"
static dispatcher **dispatchers;
static char debug = 0;
static pthread_t collectorid;
static char keep_running = 1;
int collector_interval = 60;
static char cluster_refresh_pending = 0;
static router *pending_router = NULL;
/**
* Collects metrics from dispatchers and servers and emits them.
*/
static void *
collector_runner(void *s)
{
int i;
size_t totticks;
size_t totmetrics;
size_t totblackholes;
size_t totqueued;
size_t totstalls;
size_t totdropped;
size_t totsleeps;
size_t ticks;
size_t metrics;
size_t blackholes;
size_t queued;
size_t stalls;
size_t dropped;
size_t sleeps;
time_t now;
time_t nextcycle;
char destbuf[1024]; /* sort of POSIX_MAX_PATH */
char *p;
size_t numaggregators = 0;
aggregator *aggrs = NULL;
server *submission = (server *)s;
server **srvs = NULL;
char metric[METRIC_BUFSIZ];
char *m = NULL;
size_t sizem = 0;
size_t (*s_ticks)(server *);
size_t (*s_metrics)(server *);
size_t (*s_stalls)(server *);
size_t (*s_dropped)(server *);
size_t (*d_ticks)(dispatcher *);
size_t (*d_metrics)(dispatcher *);
size_t (*d_blackholes)(dispatcher *);
size_t (*d_sleeps)(dispatcher *);
size_t (*a_received)(aggregator *);
size_t (*a_sent)(aggregator *);
size_t (*a_dropped)(aggregator *);
/* setup functions to target what the user wants */
if (debug & 2) {
s_ticks = server_get_ticks_sub;
s_metrics = server_get_metrics_sub;
s_stalls = server_get_stalls_sub;
s_dropped = server_get_dropped_sub;
d_ticks = dispatch_get_ticks_sub;
d_metrics = dispatch_get_metrics_sub;
d_blackholes = dispatch_get_blackholes_sub;
d_sleeps = dispatch_get_sleeps_sub;
a_received = aggregator_get_received_sub;
a_sent = aggregator_get_sent_sub;
a_dropped = aggregator_get_dropped_sub;
} else {
s_ticks = server_get_ticks;
s_metrics = server_get_metrics;
s_stalls = server_get_stalls;
s_dropped = server_get_dropped;
d_ticks = dispatch_get_ticks;
d_metrics = dispatch_get_metrics;
d_blackholes = dispatch_get_blackholes;
d_sleeps = dispatch_get_sleeps;
a_received = aggregator_get_received;
a_sent = aggregator_get_sent;
a_dropped = aggregator_get_dropped;
}
#define send(metric) \
if (debug & 1) \
logout("%s", metric); \
else \
server_send(submission, strdup(metric), 1);
nextcycle = time(NULL) + collector_interval;
while (keep_running) {
if (cluster_refresh_pending) {
char *stub = router_getcollectorstub(pending_router);
server **newservers = router_getservers(pending_router);
if (srvs != NULL)
free(srvs);
srvs = newservers;
aggrs = router_getaggregators(pending_router);
numaggregators = aggregator_numaggregators(aggrs);
/* prepare hostname for graphite metrics */
snprintf(metric, sizeof(metric), "%scarbon.relays.%s",
stub == NULL ? "" : stub, relay_hostname);
for (m = metric + strlen("carbon.relays."); *m != '\0'; m++)
if (*m == '.')
*m = '_';
*m++ = '.';
*m = '\0';
sizem = sizeof(metric) - (m - metric);
cluster_refresh_pending = 0;
}
assert(srvs != NULL);
sleep(1);
now = time(NULL);
if (nextcycle > now)
continue;
nextcycle += collector_interval;
totticks = 0;
totmetrics = 0;
totblackholes = 0;
totsleeps = 0;
for (i = 0; dispatchers[i] != NULL; i++) {
totsleeps += sleeps = d_sleeps(dispatchers[i]);
totticks += ticks = d_ticks(dispatchers[i]);
totmetrics += metrics = d_metrics(dispatchers[i]);
totblackholes += blackholes = d_blackholes(dispatchers[i]);
snprintf(m, sizem, "dispatcher%d.metricsReceived %zu %zu\n",
i + 1, metrics, (size_t)now);
send(metric);
snprintf(m, sizem, "dispatcher%d.metricsBlackholed %zu %zu\n",
i + 1, blackholes, (size_t)now);
send(metric);
snprintf(m, sizem, "dispatcher%d.wallTime_us %zu %zu\n",
i + 1, ticks, (size_t)now);
send(metric);
snprintf(m, sizem, "dispatcher%d.sleepTime_us %zu %zu\n",
i + 1, sleeps, (size_t)now);
send(metric);
}
snprintf(m, sizem, "metricsReceived %zu %zu\n",
totmetrics, (size_t)now);
send(metric);
snprintf(m, sizem, "metricsBlackholed %zu %zu\n",
totblackholes, (size_t)now);
send(metric);
snprintf(m, sizem, "dispatch_wallTime_us %zu %zu\n",
totticks, (size_t)now);
send(metric);
snprintf(m, sizem, "dispatch_sleepTime_us %zu %zu\n",
totsleeps, (size_t)now);
send(metric);
#define send_server_metrics(ipbuf, ticks, metrics, queued, stalls, dropped) \
snprintf(m, sizem, "destinations.%s.sent %zu %zu\n", \
ipbuf, metrics, (size_t)now); \
send(metric); \
snprintf(m, sizem, "destinations.%s.queued %zu %zu\n", \
ipbuf, queued, (size_t)now); \
send(metric); \
snprintf(m, sizem, "destinations.%s.stalls %zu %zu\n", \
ipbuf, stalls, (size_t)now); \
send(metric); \
snprintf(m, sizem, "destinations.%s.dropped %zu %zu\n", \
ipbuf, dropped, (size_t)now); \
send(metric); \
snprintf(m, sizem, "destinations.%s.wallTime_us %zu %zu\n", \
ipbuf, ticks, (size_t)now); \
send(metric);
totticks = 0;
totmetrics = 0;
totqueued = 0;
totstalls = 0;
totdropped = 0;
/* exclude internal_submission metrics from the totals to avoid
* artificial doubles due to internal routing details */
ticks = s_ticks(submission);
metrics = s_metrics(submission);
queued = server_get_queue_len(submission);
stalls = s_stalls(submission);
dropped = s_dropped(submission);
send_server_metrics(server_ip(submission),
ticks, metrics, queued, stalls, dropped);
for (i = 0; srvs[i] != NULL; i++) {
switch (server_ctype(srvs[i])) {
case CON_FILE:
case CON_PIPE:
snprintf(destbuf, sizeof(destbuf), "%s",
server_ip(srvs[i]));
break;
case CON_TCP:
snprintf(destbuf, sizeof(destbuf), "%s:%u",
server_ip(srvs[i]), server_port(srvs[i]));
break;
case CON_UDP:
snprintf(destbuf, sizeof(destbuf), "%s:%u-udp",
server_ip(srvs[i]), server_port(srvs[i]));
break;
}
for (p = destbuf; *p != '\0'; p++)
if (*p == '.')
*p = '_';
totticks += ticks = s_ticks(srvs[i]);
totmetrics += metrics = s_metrics(srvs[i]);
totqueued += queued = server_get_queue_len(srvs[i]);
totstalls += stalls = s_stalls(srvs[i]);
totdropped += dropped = s_dropped(srvs[i]);
send_server_metrics(destbuf,
ticks, metrics, queued, stalls, dropped);
}
snprintf(m, sizem, "metricsSent %zu %zu\n",
totmetrics, (size_t)now);
send(metric);
snprintf(m, sizem, "metricsQueued %zu %zu\n",
totqueued, (size_t)now);
send(metric);
snprintf(m, sizem, "metricStalls %zu %zu\n",
totstalls, (size_t)now);
send(metric);
snprintf(m, sizem, "metricsDropped %zu %zu\n",
totdropped, (size_t)now);
send(metric);
snprintf(m, sizem, "server_wallTime_us %zu %zu\n",
totticks, (size_t)now);
send(metric);
snprintf(m, sizem, "connections %zu %zu\n",
dispatch_get_accepted_connections(), (size_t)now);
send(metric);
snprintf(m, sizem, "disconnects %zu %zu\n",
dispatch_get_closed_connections(), (size_t)now);
send(metric);
if (numaggregators > 0) {
snprintf(m, sizem, "aggregators.metricsReceived %zu %zu\n",
a_received(aggrs), (size_t)now);
send(metric);
snprintf(m, sizem, "aggregators.metricsSent %zu %zu\n",
a_sent(aggrs), (size_t)now);
send(metric);
snprintf(m, sizem, "aggregators.metricsDropped %zu %zu\n",
a_dropped(aggrs), (size_t)now);
send(metric);
}
if (debug & 1)
fflush(stdout);
}
if (srvs != NULL)
free(srvs);
return NULL;
}
/**
* Writes messages about dropped events or high queue sizes.
*/
static size_t lastdropped = 0;
static size_t lastaggrdropped = 0;
static void *
collector_writer(void *unused)
{
int i = 0;
size_t queued;
size_t queuesize;
double queueusage;
size_t totdropped;
size_t lastconn = 0;
size_t lastdisc = 0;
size_t numaggregators = 0;
server **srvs = NULL;
aggregator *aggrs = NULL;
while (keep_running) {
if (cluster_refresh_pending) {
server **newservers = router_getservers(pending_router);
if (srvs != NULL)
free(srvs);
srvs = newservers;
aggrs = router_getaggregators(pending_router);
numaggregators = aggregator_numaggregators(aggrs);
cluster_refresh_pending = 0;
}
assert(srvs != NULL);
sleep(1);
if (debug & 1) {
size_t mpsout;
size_t totout;
size_t mpsdrop;
size_t totdrop;
size_t totqueue;
size_t mpsin;
size_t totin;
size_t totconn;
size_t totdisc;
size_t dticks;
size_t dsleeps;
int j;
/* Solaris iostat like output:
metrics in metrics out metrics drop queue conns disconn workr
mps tot mps tot dps tot cur cps tot dps tot act
99999 9999999 99999 9999999 99999 9999999 99999 999 99999 999 99999 99%
*/
if (i % 24 == 0)
printf(" metrics in metrics out metrics drop queue conns disconn workr\n"
" mps tot mps tot dps tot cur cps tot dps tot act\n");
mpsout = totout = 0;
mpsdrop = totdrop = 0;
totqueue = 0;
for (j = 0; srvs[j] != NULL; j++) {
mpsout += server_get_metrics_sub(srvs[j]);
totout += server_get_metrics(srvs[j]);
mpsdrop += server_get_dropped_sub(srvs[j]);
totdrop += server_get_dropped(srvs[j]);
totqueue += server_get_queue_len(srvs[j]);
}
mpsin = totin = 0;
dticks = dsleeps = 0;
for (j = 0; dispatchers[j] != NULL; j++) {
mpsin += dispatch_get_metrics_sub(dispatchers[j]);
totin += dispatch_get_metrics(dispatchers[j]);
dticks += dispatch_get_ticks_sub(dispatchers[j]);
dsleeps += dispatch_get_sleeps_sub(dispatchers[j]);
}
totconn = dispatch_get_accepted_connections();
totdisc = dispatch_get_closed_connections();
printf("%5zu %7zu " /* metrics in */
"%5zu %7zu " /* metrics out */
"%5zu %7zu " /* metrics dropped */
"%5zu " /* queue */
"%3zu %5zu " /* conns */
"%3zu %5zu " /* disconns */
"%2d%%\n", /* workers */
mpsin, totin,
mpsout, totout,
mpsdrop, totdrop,
totqueue,
totconn - lastconn, totconn,
totdisc - lastdisc, totdisc,
(int)(((double)dticks * 100.0) / (double)(dticks + dsleeps))
);
lastconn = totconn;
lastdisc = totdisc;
}
i++;
if (i < collector_interval)
continue;
totdropped = 0;
for (i = 0; srvs[i] != NULL; i++) {
queued = server_get_queue_len(srvs[i]);
queuesize = server_get_queue_size(srvs[i]);
totdropped += server_get_dropped(srvs[i]);
queueusage = (double)queued / (double)queuesize;
if (queueusage >= 0.75)
logout("warning: metrics queuing up "
"for %s:%u: %zu metrics (%d%% of queue size)\n",
server_ip(srvs[i]), server_port(srvs[i]), queued,
(int)(queueusage * 100));
}
if (totdropped - lastdropped > 0)
logout("warning: dropped %zu metrics\n", totdropped - lastdropped);
lastdropped = totdropped;
if (numaggregators > 0) {
totdropped = aggregator_get_dropped(aggrs);
if (totdropped - lastaggrdropped > 0)
logout("warning: aggregator dropped %zu metrics\n",
totdropped - lastaggrdropped);
lastaggrdropped = totdropped;
}
i = 0;
}
return NULL;
}
/**
* Schedules routes r to be put in place for the current routes. The
* replacement is performed at the next cycle of the collector.
*/
inline void
collector_schedulereload(router *rtr)
{
pending_router = rtr;
cluster_refresh_pending = 1;
}
/**
* Returns true if the routes scheduled to be reloaded by a call to
* collector_schedulereload() have been activated.
*/
inline char
collector_reloadcomplete(void)
{
return cluster_refresh_pending == 0;
}
/**
* Initialises and starts the collector.
*/
void
collector_start(dispatcher **d, router *rtr, server *submission, char cum)
{
dispatchers = d;
collector_schedulereload(rtr);
if (mode & MODE_DEBUG)
debug = 1;
debug |= (cum ? 0 : 2);
if (mode & MODE_SUBMISSION) {
if (pthread_create(&collectorid, NULL, collector_writer, NULL) != 0)
logerr("failed to start collector!\n");
} else {
if (pthread_create(&collectorid, NULL, collector_runner, submission) != 0)
logerr("failed to start collector!\n");
}
}
/**
* Shuts down the collector.
*/
void
collector_stop(void)
{
keep_running = 0;
pthread_join(collectorid, NULL);
}
|