forked from envoyproxy/nighthawk
-
Notifications
You must be signed in to change notification settings - Fork 0
/
statistic_test.cc
478 lines (408 loc) · 17.3 KB
/
statistic_test.cc
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
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
#include <google/protobuf/util/json_util.h>
#include <chrono>
#include <random>
#include <string>
#include <typeinfo> // std::bad_cast
#include "external/envoy/source/common/protobuf/utility.h"
#include "external/envoy/source/common/stats/isolated_store_impl.h"
#include "external/envoy/test/mocks/stats/mocks.h"
#include "external/envoy/test/test_common/file_system_for_test.h"
#include "external/envoy/test/test_common/utility.h"
#include "source/common/statistic_impl.h"
#include "test/test_common/environment.h"
#include "gtest/gtest.h"
using namespace std::chrono_literals;
using namespace testing;
namespace Nighthawk {
using MyTypes = Types<SimpleStatistic, InMemoryStatistic, HdrStatistic, StreamingStatistic,
CircllhistStatistic>;
template <typename T> class TypedStatisticTest : public Test {};
class Helper {
public:
/**
* With 0 significant digits passed, this uses EXPECT_DOUBLE_EQ. Otherwise expectNear
* will be called with a computed acceptable range based on the number of significant
* digits and tested_value.
* @param expected_value the expected value
* @param tested_value the tested_value
* @param significant the number of significant digits that should be used to compare values.
*/
static void expectNear(double expected_value, double tested_value, uint64_t significant) {
if (significant > 0) {
EXPECT_NEAR(expected_value, tested_value,
std::pow(10, std::ceil(std::log10(tested_value)) - 1 - significant));
} else {
EXPECT_DOUBLE_EQ(expected_value, tested_value);
}
}
};
TYPED_TEST_SUITE(TypedStatisticTest, MyTypes);
TYPED_TEST(TypedStatisticTest, Simple) {
TypeParam a;
TypeParam b;
std::vector<int> a_values{1, 2, 3};
std::vector<int> b_values{1234, 6543456, 342335};
for (int value : a_values) {
a.addValue(value);
}
EXPECT_EQ(3, a.count());
EXPECT_EQ(1, a.min());
EXPECT_EQ(3, a.max());
for (int value : b_values) {
b.addValue(value);
}
EXPECT_EQ(3, b.count());
EXPECT_EQ(1234, b.min());
// We substract one from the expected precision with respect to significant digits for
// HdrHistogram. (More context in comments over at the the HdrStatisticProtoOutputLargeValues test
// below).
Helper::expectNear(6543456, b.max(), b.significantDigits() - 1);
Helper::expectNear(2.0, a.mean(), a.significantDigits());
Helper::expectNear(0.6666666666666666, a.pvariance(), a.significantDigits());
Helper::expectNear(0.816496580927726, a.pstdev(), a.significantDigits());
Helper::expectNear(2295675.0, b.mean(), a.significantDigits());
Helper::expectNear(9041213360680.666, b.pvariance(), a.significantDigits());
Helper::expectNear(3006861.0477839955, b.pstdev(), a.significantDigits());
auto c = a.combine(b);
EXPECT_EQ(6, c->count());
EXPECT_EQ(1, c->min());
Helper::expectNear(6543456, c->max(), c->significantDigits() - 1);
Helper::expectNear(1147838.5, c->mean(), c->significantDigits());
Helper::expectNear(5838135311072.917, c->pvariance(), c->significantDigits());
Helper::expectNear(2416223.357033227, c->pstdev(), c->significantDigits());
// A reverse combine should be exactly equivalent.
auto d = b.combine(a);
EXPECT_EQ(c->count(), d->count());
EXPECT_EQ(c->min(), d->min());
EXPECT_EQ(c->max(), d->max());
EXPECT_EQ(c->mean(), d->mean());
EXPECT_EQ(c->pvariance(), d->pvariance());
EXPECT_EQ(c->pstdev(), d->pstdev());
}
TYPED_TEST(TypedStatisticTest, createNewInstanceOfSameType) {
TypeParam a;
EXPECT_NE(a.createNewInstanceOfSameType(), nullptr);
}
TYPED_TEST(TypedStatisticTest, Empty) {
TypeParam a;
EXPECT_EQ(0, a.count());
EXPECT_TRUE(std::isnan(a.mean()));
EXPECT_TRUE(std::isnan(a.pvariance()));
EXPECT_TRUE(std::isnan(a.pstdev()));
EXPECT_EQ(a.min(), UINT64_MAX);
EXPECT_EQ(a.max(), 0);
}
TYPED_TEST(TypedStatisticTest, SingleAndDoubleValue) {
TypeParam a;
a.addValue(1);
EXPECT_EQ(1, a.count());
Helper::expectNear(1.0, a.mean(), a.significantDigits());
EXPECT_DOUBLE_EQ(0, a.pvariance());
EXPECT_DOUBLE_EQ(0, a.pstdev());
a.addValue(2);
EXPECT_EQ(2, a.count());
Helper::expectNear(1.5, a.mean(), a.significantDigits());
Helper::expectNear(0.25, a.pvariance(), a.significantDigits());
Helper::expectNear(0.5, a.pstdev(), a.significantDigits());
}
TYPED_TEST(TypedStatisticTest, CatastrophicalCancellation) {
// From https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance
// Assume that all floating point operations use standard IEEE 754 double-precision arithmetic.
// Consider the sample (4, 7, 13, 16) from an infinite population. Based on this sample, the
// estimated population mean is 10, and the unbiased estimate of population variance is 30. Both
// the naïve algorithm and two-pass algorithm compute these values correctly.
// Next consider the sample (108 + 4, 108 + 7, 108 + 13, 108 + 16), which gives rise to the same
// estimated variance as the first sample. The two-pass algorithm computes this variance estimate
// correctly, but the naïve algorithm returns 29.333333333333332 instead of 30. While this loss of
// precision may be tolerable and viewed as a minor flaw of the naïve algorithm, further
// increasing the offset makes the error catastrophic. Consider the sample (109 + 4, 109 + 7, 109
// + 13, 109 + 16). Again the estimated population variance of 30 is computed correctly by the
// two-pass algorithm, but the naïve algorithm now computes it as −170.66666666666666. This is a
// serious problem with naïve algorithm and is due to catastrophic cancellation in the subtraction
// of two similar numbers at the final stage of the algorithm.
std::vector<uint64_t> values{4, 7, 13, 16};
uint64_t exponential = 0;
for (exponential = 3; exponential < 16; exponential++) {
TypeParam a;
double offset = std::pow(10ULL, exponential);
for (int value : values) {
a.addValue(offset + value);
}
// If an implementation makes this claim, we put it to the test. SimpleStatistic is simple and
// fast, but starts failing this test when exponential equals 8. HdrStatistic breaks at 5.
// TODO(oschaaf): evaluate ^^
if (a.resistsCatastrophicCancellation()) {
Helper::expectNear(22.5, a.pvariance(), a.significantDigits());
Helper::expectNear(4.7434164902525691, a.pstdev(), a.significantDigits());
}
}
}
TYPED_TEST(TypedStatisticTest, OneMillionRandomSamples) {
std::mt19937_64 mt(1243);
// TODO(oschaaf): Actually the range we want to test is a factor 1000 higher, but
// then catastrophical cancellation make SimpleStatistic fail expectations.
// For now, we use values that shouldn't trigger the phenomena. Revisit this later.
std::uniform_int_distribution<uint64_t> dist(1ULL, 1000ULL * 1000 * 60);
StreamingStatistic referenceStatistic;
TypeParam testStatistic;
for (int i = 0; i < 999999; ++i) {
auto value = dist(mt);
referenceStatistic.addValue(value);
testStatistic.addValue(value);
}
Helper::expectNear(referenceStatistic.mean(), testStatistic.mean(),
testStatistic.significantDigits());
Helper::expectNear(referenceStatistic.pvariance(), testStatistic.pvariance(),
testStatistic.significantDigits());
Helper::expectNear(referenceStatistic.pstdev(), testStatistic.pstdev(),
testStatistic.significantDigits());
}
TYPED_TEST(TypedStatisticTest, ProtoOutput) {
TypeParam a;
a.setId("foo");
a.addValue(6543456);
a.addValue(342335);
const nighthawk::client::Statistic proto = a.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ("foo", proto.id());
EXPECT_EQ(2, proto.count());
EXPECT_EQ(std::round(a.mean()), proto.mean().nanos());
EXPECT_EQ(std::round(a.pstdev()), proto.pstdev().nanos());
}
TYPED_TEST(TypedStatisticTest, ProtoOutputEmptyStats) {
TypeParam a;
const nighthawk::client::Statistic proto = a.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ(proto.count(), 0);
EXPECT_EQ(proto.mean().nanos(), 0);
EXPECT_EQ(proto.pstdev().nanos(), 0);
}
TYPED_TEST(TypedStatisticTest, NativeRoundtrip) {
TypeParam a;
a.setId("bar");
a.addValue(6543456);
a.addValue(342335);
a.addValue(543);
const absl::StatusOr<std::unique_ptr<std::istream>> status_or_stream = a.serializeNative();
if (status_or_stream.ok()) {
// If the histogram states it implements native serialization/deserialization, put it through
// a round trip test.
TypeParam b;
absl::Status status = b.deserializeNative(*status_or_stream.value());
EXPECT_TRUE(status.ok());
EXPECT_EQ(3, b.count());
EXPECT_EQ(a.count(), b.count());
EXPECT_EQ(a.mean(), b.mean());
EXPECT_EQ(a.pstdev(), b.pstdev());
} else {
EXPECT_EQ(status_or_stream.status().code(), absl::StatusCode::kUnimplemented);
}
}
TYPED_TEST(TypedStatisticTest, AttemptsToDeserializeBogusBehaveWell) {
// Deserializing corrupted data should either result in the statistic reporting
// it didn't implement deserialization, or having it report an internal failure.
const std::vector<absl::StatusCode> expected_status_list{absl::StatusCode::kInternal,
absl::StatusCode::kUnimplemented};
TypeParam a;
std::istringstream bogus_input(std::string("BOGUS"));
const absl::Status status = a.deserializeNative(bogus_input);
EXPECT_FALSE(status.ok());
EXPECT_THAT(expected_status_list, Contains(status.code()));
}
TYPED_TEST(TypedStatisticTest, StringOutput) {
TypeParam a;
a.addValue(6543456);
a.addValue(342335);
std::string s = a.toString();
std::vector<std::string> matches{
"count: ", "raw_mean: ", "raw_pstdev: ", "raw_min: ", "raw_max: "};
for (const auto& match : matches) {
EXPECT_NE(std::string::npos, s.find(match));
}
}
TYPED_TEST(TypedStatisticTest, IdFieldWorks) {
TypeParam statistic;
std::string id = "fooid";
EXPECT_EQ("", statistic.id());
statistic.setId(id);
EXPECT_EQ(id, statistic.id());
}
class StatisticTest : public Test {};
// Note that we explicitly subject SimpleStatistic to the large
// values below, and see a 0 stdev returned.
TEST(StatisticTest, SimpleStatisticProtoOutputLargeValues) {
SimpleStatistic a;
uint64_t value = 100ul + 0xFFFFFFFF; // 100 + the max for uint32_t
a.addValue(value);
a.addValue(value);
const nighthawk::client::Statistic proto = a.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ(proto.count(), 2);
Helper::expectNear(((1.0 * proto.mean().seconds() * 1000 * 1000 * 1000) + proto.mean().nanos()),
value, a.significantDigits() - 1);
// 0 because std::nan() gets translated to that.
EXPECT_EQ(proto.pstdev().nanos(), 0);
}
TEST(StatisticTest, HdrStatisticProtoOutputLargeValues) {
HdrStatistic a;
uint64_t value = 100ul + 0xFFFFFFFF;
a.addValue(value);
a.addValue(value);
const nighthawk::client::Statistic proto = a.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ(proto.count(), 2);
// TODO(oschaaf): hdr doesn't seem to the promised precision in this scenario.
// We substract one from the indicated significant digits to make this test pass.
// TODO(oschaaf): revisit this to make sure there's not a different underlying problem.
Helper::expectNear(((1.0 * proto.mean().seconds() * 1000 * 1000 * 1000) + proto.mean().nanos()),
value, a.significantDigits() - 1);
EXPECT_EQ(proto.pstdev().nanos(), 0);
}
TEST(StatisticTest, StreamingStatProtoOutputLargeValues) {
StreamingStatistic a;
uint64_t value = 100ul + 0xFFFFFFFF;
a.addValue(value);
a.addValue(value);
const nighthawk::client::Statistic proto = a.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ(proto.count(), 2);
Helper::expectNear(((1.0 * proto.mean().seconds() * 1000 * 1000 * 1000) + proto.mean().nanos()),
value, a.significantDigits());
EXPECT_EQ(proto.pstdev().nanos(), 0);
}
TEST(StatisticTest, CircllhistStatisticProtoOutputLargeValues) {
CircllhistStatistic statistic;
uint64_t value = 100ul + 0xFFFFFFFF;
statistic.addValue(value);
statistic.addValue(value);
const nighthawk::client::Statistic proto =
statistic.toProto(Statistic::SerializationDomain::DURATION);
EXPECT_EQ(proto.count(), 2);
Helper::expectNear(Envoy::Protobuf::util::TimeUtil::DurationToNanoseconds(proto.mean()), value,
statistic.significantDigits());
EXPECT_EQ(Envoy::Protobuf::util::TimeUtil::DurationToNanoseconds(proto.pstdev()), 0);
}
TEST(StatisticTest, HdrStatisticPercentilesProto) {
nighthawk::client::Statistic parsed_json_proto;
HdrStatistic statistic;
for (int i = 1; i <= 10; i++) {
statistic.addValue(i);
}
Envoy::MessageUtil util;
util.loadFromJson(
Envoy::Filesystem::fileSystemForTest()
.fileReadToEnd(TestEnvironment::runfilesPath("test/test_data/hdr_proto_json.gold"))
.value(),
parsed_json_proto, Envoy::ProtobufMessage::getStrictValidationVisitor());
const std::string json = util.getJsonStringFromMessageOrError(
statistic.toProto(Statistic::SerializationDomain::DURATION), true, true);
const std::string golden_json =
util.getJsonStringFromMessageOrError(parsed_json_proto, true, true);
EXPECT_THAT(statistic.toProto(Statistic::SerializationDomain::DURATION),
Envoy::ProtoEq(parsed_json_proto))
<< json << "\n"
<< "is not equal to golden file:\n"
<< golden_json;
}
TEST(StatisticTest, CircllhistStatisticPercentilesProto) {
nighthawk::client::Statistic parsed_json_proto;
CircllhistStatistic statistic;
for (int i = 1; i <= 10; i++) {
statistic.addValue(i);
}
Envoy::MessageUtil util;
util.loadFromJson(
Envoy::Filesystem::fileSystemForTest()
.fileReadToEnd(TestEnvironment::runfilesPath("test/test_data/circllhist_proto_json.gold"))
.value(),
parsed_json_proto, Envoy::ProtobufMessage::getStrictValidationVisitor());
const std::string json = util.getJsonStringFromMessageOrError(
statistic.toProto(Statistic::SerializationDomain::DURATION), true, true);
const std::string golden_json =
util.getJsonStringFromMessageOrError(parsed_json_proto, true, true);
EXPECT_THAT(statistic.toProto(Statistic::SerializationDomain::DURATION),
Envoy::ProtoEq(parsed_json_proto))
<< json << "\n"
<< "is not equal to golden file:\n"
<< golden_json;
}
TEST(StatisticTest, CombineAcrossTypesFails) {
HdrStatistic a;
InMemoryStatistic b;
StreamingStatistic c;
CircllhistStatistic d;
EXPECT_THROW(a.combine(b), std::bad_cast);
EXPECT_THROW(a.combine(c), std::bad_cast);
EXPECT_THROW(b.combine(a), std::bad_cast);
EXPECT_THROW(b.combine(c), std::bad_cast);
EXPECT_THROW(c.combine(a), std::bad_cast);
EXPECT_THROW(c.combine(b), std::bad_cast);
EXPECT_THROW(c.combine(d), std::bad_cast);
EXPECT_THROW(d.combine(a), std::bad_cast);
}
TEST(StatisticTest, HdrStatisticOutOfRange) {
HdrStatistic a;
a.addValue(INT64_MAX);
EXPECT_EQ(0, a.count());
}
TEST(StatisticTest, NullStatistic) {
NullStatistic stat;
EXPECT_EQ(0, stat.count());
std::string id = "fooid";
stat.setId(id);
EXPECT_EQ(id, stat.id());
stat.addValue(1);
EXPECT_EQ(0, stat.count());
EXPECT_EQ(0, stat.max());
EXPECT_EQ(UINT64_MAX, stat.min());
EXPECT_EQ(0, stat.mean());
EXPECT_EQ(0, stat.pvariance());
EXPECT_EQ(0, stat.pstdev());
EXPECT_NE(nullptr, stat.combine(stat));
EXPECT_EQ(0, stat.significantDigits());
EXPECT_NE(nullptr, stat.createNewInstanceOfSameType());
const nighthawk::client::Statistic proto = stat.toProto(Statistic::SerializationDomain::RAW);
EXPECT_EQ(id, proto.id());
EXPECT_EQ(0, proto.count());
EXPECT_EQ(0, proto.raw_mean());
EXPECT_EQ(0, proto.raw_pstdev());
EXPECT_EQ(0, proto.raw_max());
EXPECT_EQ(UINT64_MAX, proto.raw_min());
}
using SinkableTypes = Types<SinkableHdrStatistic, SinkableCircllhistStatistic>;
template <typename T> class SinkableStatisticTest : public Test {};
TYPED_TEST_SUITE(SinkableStatisticTest, SinkableTypes);
TYPED_TEST(SinkableStatisticTest, EmptySinkableStatistic) {
Envoy::Stats::MockIsolatedStatsStore mock_store;
TypeParam stat(*mock_store.rootScope());
EXPECT_EQ(0, stat.count());
EXPECT_TRUE(std::isnan(stat.mean()));
EXPECT_TRUE(std::isnan(stat.pvariance()));
EXPECT_TRUE(std::isnan(stat.pstdev()));
EXPECT_EQ(stat.min(), UINT64_MAX);
EXPECT_EQ(stat.max(), 0);
EXPECT_EQ(Envoy::Stats::Histogram::Unit::Unspecified, stat.unit());
EXPECT_FALSE(stat.used());
EXPECT_EQ("", stat.name());
EXPECT_EQ("", stat.tagExtractedName());
EXPECT_EQ(absl::nullopt, stat.worker_id());
}
TYPED_TEST(SinkableStatisticTest, SimpleSinkableStatistic) {
Envoy::Stats::MockIsolatedStatsStore mock_store;
const int worker_id = 0;
TypeParam stat(*mock_store.rootScope(), worker_id);
const uint64_t sample_value = 123;
const std::string stat_name = "stat_name";
EXPECT_CALL(mock_store, deliverHistogramToSinks(_, sample_value)).Times(2);
stat.recordValue(sample_value);
stat.addValue(sample_value);
stat.setId(stat_name);
EXPECT_EQ(2, stat.count());
Helper::expectNear(123.0, stat.mean(), stat.significantDigits());
EXPECT_DOUBLE_EQ(0, stat.pvariance());
EXPECT_DOUBLE_EQ(0, stat.pstdev());
EXPECT_EQ(123, stat.min());
EXPECT_EQ(123, stat.max());
EXPECT_EQ(Envoy::Stats::Histogram::Unit::Unspecified, stat.unit());
EXPECT_TRUE(stat.used());
EXPECT_EQ(stat_name, stat.name());
EXPECT_EQ("0.stat_name", stat.tagExtractedName());
EXPECT_TRUE(stat.worker_id().has_value());
EXPECT_EQ(worker_id, stat.worker_id().value());
}
} // namespace Nighthawk