Grcov report - health_checker_base

1

#include "source/extensions/health_checkers/common/health_checker_base_impl.h"

2

3

#include "envoy/config/core/v3/address.pb.h"

4

#include "envoy/config/core/v3/health_check.pb.h"

5

#include "envoy/data/core/v3/health_check_event.pb.h"

6

#include "envoy/stats/scope.h"

7

8

#include "source/common/network/utility.h"

9

#include "source/common/router/router.h"

10

11

namespace Envoy {

12

namespace Upstream {

13

14

HealthCheckerImplBase::HealthCheckerImplBase(const Cluster& cluster,

15

                                             const envoy::config::core::v3::HealthCheck& config,

16

                                             Event::Dispatcher& dispatcher,

17

                                             Runtime::Loader& runtime,

18

                                             Random::RandomGenerator& random,

19

                                             HealthCheckEventLoggerPtr&& event_logger)

20

359

    : always_log_health_check_failures_(config.always_log_health_check_failures()),

21

359

      always_log_health_check_success_(config.always_log_health_check_success()), cluster_(cluster),

22

359

      dispatcher_(dispatcher), timeout_(PROTOBUF_GET_MS_REQUIRED(config, timeout)),

23

359

      unhealthy_threshold_(PROTOBUF_GET_WRAPPED_REQUIRED(config, unhealthy_threshold)),

24

359

      healthy_threshold_(PROTOBUF_GET_WRAPPED_REQUIRED(config, healthy_threshold)),

25

359

      stats_(generateStats(cluster.info()->statsScope())), runtime_(runtime), random_(random),

26

359

      reuse_connection_(PROTOBUF_GET_WRAPPED_OR_DEFAULT(config, reuse_connection, true)),

27

359

      event_logger_(std::move(event_logger)), interval_(PROTOBUF_GET_MS_REQUIRED(config, interval)),

28

359

      no_traffic_interval_(PROTOBUF_GET_MS_OR_DEFAULT(config, no_traffic_interval, 60000)),

29

359

      no_traffic_healthy_interval_(PROTOBUF_GET_MS_OR_DEFAULT(config, no_traffic_healthy_interval,

30

359

                                                              no_traffic_interval_.count())),

31

359

      initial_jitter_(PROTOBUF_GET_MS_OR_DEFAULT(config, initial_jitter, 0)),

32

359

      interval_jitter_(PROTOBUF_GET_MS_OR_DEFAULT(config, interval_jitter, 0)),

33

359

      interval_jitter_percent_(config.interval_jitter_percent()),

34

359

      unhealthy_interval_(

35

359

          PROTOBUF_GET_MS_OR_DEFAULT(config, unhealthy_interval, interval_.count())),

36

359

      unhealthy_edge_interval_(

37

359

          PROTOBUF_GET_MS_OR_DEFAULT(config, unhealthy_edge_interval, unhealthy_interval_.count())),

38

359

      healthy_edge_interval_(

39

359

          PROTOBUF_GET_MS_OR_DEFAULT(config, healthy_edge_interval, interval_.count())),

40

359

      transport_socket_options_(initTransportSocketOptions(config)),

41

359

      transport_socket_match_metadata_(initTransportSocketMatchMetadata(config)),

42

359

      member_update_cb_{cluster_.prioritySet().addMemberUpdateCb(

43

505

          [this](const HostVector& hosts_added, const HostVector& hosts_removed) {

44

251

            onClusterMemberUpdate(hosts_added, hosts_removed);

45

505

          })} {}

46

47

std::shared_ptr<const Network::TransportSocketOptionsImpl>

48

HealthCheckerImplBase::initTransportSocketOptions(

49

359

    const envoy::config::core::v3::HealthCheck& config) {

50

359

  if (config.has_tls_options()) {

51

1

    std::vector<std::string> protocols{config.tls_options().alpn_protocols().begin(),

52

1

                                       config.tls_options().alpn_protocols().end()};

53

1

    return std::make_shared<const Network::TransportSocketOptionsImpl>(

54

1

        "", std::vector<std::string>{}, std::move(protocols));

55

1

56

57

358

  return std::make_shared<const Network::TransportSocketOptionsImpl>();

58

359

59

60

MetadataConstSharedPtr HealthCheckerImplBase::initTransportSocketMatchMetadata(

61

359

    const envoy::config::core::v3::HealthCheck& config) {

62

359

  if (config.has_transport_socket_match_criteria()) {

63

8

    std::shared_ptr<envoy::config::core::v3::Metadata> metadata =

64

8

        std::make_shared<envoy::config::core::v3::Metadata>();

65

8

    (*metadata->mutable_filter_metadata())[Envoy::Config::MetadataFilters::get()

66

8

                                               .ENVOY_TRANSPORT_SOCKET_MATCH] =

67

8

        config.transport_socket_match_criteria();

68

8

    return metadata;

69

8

70

71

351

  return nullptr;

72

359

73

74

359

HealthCheckerImplBase::~HealthCheckerImplBase() {

75

  // First clear callbacks that otherwise will be run from

76

  // ActiveHealthCheckSession::onDeferredDeleteBase(). This prevents invoking a callback on a

77

  // deleted parent object (e.g. Cluster).

78

359

  callbacks_.clear();

79

  // ASSERTs inside the session destructor check to make sure we have been previously deferred

80

  // deleted. Unify that logic here before actual destruction happens.

81

372

  for (auto& session : active_sessions_) {

82

326

    session.second->onDeferredDeleteBase();

83

326

84

359

85

86

265

void HealthCheckerImplBase::decHealthy() { stats_.healthy_.sub(1); }

87

88

void HealthCheckerImplBase::decDegraded() { stats_.degraded_.sub(1); }

89

90

359

HealthCheckerStats HealthCheckerImplBase::generateStats(Stats::Scope& scope) {

91

359

  std::string prefix("health_check.");

92

359

  return {ALL_HEALTH_CHECKER_STATS(POOL_COUNTER_PREFIX(scope, prefix),

93

359

                                   POOL_GAUGE_PREFIX(scope, prefix))};

94

359

95

96

265

void HealthCheckerImplBase::incHealthy() { stats_.healthy_.add(1); }

97

98

1

void HealthCheckerImplBase::incDegraded() { stats_.degraded_.add(1); }

99

100

std::chrono::milliseconds HealthCheckerImplBase::interval(HealthState state,

101

1073

                                                          HealthTransition changed_state) const {

102

  // See if the cluster has ever made a connection. If not, we use a much slower interval to keep

103

  // the host info relatively up to date in case we suddenly start sending traffic to this cluster.

104

  // In general host updates are rare and this should greatly smooth out needless health checking.

105

  // If a connection has been established, we choose an interval based on the host's health. Please

106

  // refer to the HealthCheck API documentation for more details.

107

1073

  uint64_t base_time_ms;

108

1073

  if (cluster_.info()->trafficStats()->upstream_cx_total_.used()) {

109

    // When healthy/unhealthy threshold is configured the health transition of a host will be

110

    // delayed. In this situation Envoy should use the edge interval settings between health checks.

111

//

112

    // Example scenario for an unhealthy host with healthy_threshold set to 3:

113

    // - check fails, host is still unhealthy and next check happens after unhealthy_interval;

114

    // - check succeeds, host is still unhealthy and next check happens after healthy_edge_interval;

115

    // - check succeeds, host is still unhealthy and next check happens after healthy_edge_interval;

116

    // - check succeeds, host is now healthy and next check happens after interval;

117

    // - check succeeds, host is still healthy and next check happens after interval.

118

338

    switch (state) {

119

46

    case HealthState::Unhealthy:

120

46

      base_time_ms = changed_state == HealthTransition::ChangePending

121

46

                         ? unhealthy_edge_interval_.count()

122

46

                         : unhealthy_interval_.count();

123

46

      break;

124

292

    default:

125

292

      base_time_ms = changed_state == HealthTransition::ChangePending

126

292

                         ? healthy_edge_interval_.count()

127

292

                         : interval_.count();

128

292

      break;

129

338

130

735

  } else {

131

735

    base_time_ms =

132

735

        (state == HealthState::Healthy && changed_state != HealthTransition::ChangePending)

133

735

            ? no_traffic_healthy_interval_.count()

134

735

            : no_traffic_interval_.count();

135

735

136

1073

  return intervalWithJitter(base_time_ms, interval_jitter_);

137

1073

138

139

std::chrono::milliseconds

140

HealthCheckerImplBase::intervalWithJitter(uint64_t base_time_ms,

141

1074

                                          std::chrono::milliseconds interval_jitter) const {

142

1074

  const uint64_t jitter_percent_mod = interval_jitter_percent_ * base_time_ms / 100;

143

1074

  if (jitter_percent_mod > 0) {

144

426

    base_time_ms += random_.random() % jitter_percent_mod;

145

426

146

147

1074

  if (interval_jitter.count() > 0) {

148

352

    base_time_ms += (random_.random() % interval_jitter.count());

149

352

150

151

1074

  const uint64_t min_interval = runtime_.snapshot().getInteger("health_check.min_interval", 0);

152

1074

  const uint64_t max_interval = runtime_.snapshot().getInteger(

153

1074

      "health_check.max_interval", std::numeric_limits<uint64_t>::max());

154

155

1074

  uint64_t final_ms = std::min(base_time_ms, max_interval);

156

  // We force a non-zero final MS, to prevent live lock.

157

1074

  final_ms = std::max(uint64_t(1), std::max(final_ms, min_interval));

158

1074

  return std::chrono::milliseconds(final_ms);

159

1074

160

161

589

void HealthCheckerImplBase::addHosts(const HostVector& hosts) {

162

595

  for (const HostSharedPtr& host : hosts) {

163

356

    if (host->disableActiveHealthCheck()) {

164

4

      continue;

165

4

166

352

    active_sessions_[host] = makeSession(host);

167

352

    host->setHealthChecker(

168

352

        HealthCheckHostMonitorPtr{new HealthCheckHostMonitorImpl(shared_from_this(), host)});

169

352

    active_sessions_[host]->start();

170

352

171

589

172

173

void HealthCheckerImplBase::onClusterMemberUpdate(const HostVector& hosts_added,

174

251

                                                  const HostVector& hosts_removed) {

175

251

  addHosts(hosts_added);

176

251

  for (const HostSharedPtr& host : hosts_removed) {

177

28

    if (host->disableActiveHealthCheck()) {

178

2

      continue;

179

2

180

26

    auto session_iter = active_sessions_.find(host);

181

26

    ASSERT(active_sessions_.end() != session_iter);

182

    // This deletion can happen inline in response to a host failure, so we deferred delete.

183

26

    session_iter->second->onDeferredDeleteBase();

184

26

    dispatcher_.deferredDelete(std::move(session_iter->second));

185

26

    active_sessions_.erase(session_iter);

186

26

187

251

188

189

void HealthCheckerImplBase::runCallbacks(HostSharedPtr host, HealthTransition changed_state,

190

1149

                                         HealthState current_check_result) {

191

1403

  for (const HostStatusCb& cb : callbacks_) {

192

1233

    cb(host, changed_state, current_check_result);

193

1233

194

1149

195

196

4

void HealthCheckerImplBase::HealthCheckHostMonitorImpl::setUnhealthy(UnhealthyType type) {

197

  // This is called cross thread. The cluster/health checker might already be gone.

198

4

  std::shared_ptr<HealthCheckerImplBase> health_checker = health_checker_.lock();

199

4

  if (health_checker) {

200

4

    health_checker->setUnhealthyCrossThread(host_.lock(), type);

201

4

202

4

203

204

void HealthCheckerImplBase::setUnhealthyCrossThread(const HostSharedPtr& host,

205

4

                                                    HealthCheckHostMonitor::UnhealthyType type) {

206

4

  if (type == HealthCheckHostMonitor::UnhealthyType::ImmediateHealthCheckFail) {

207

4

    host->healthFlagSet(Host::HealthFlag::EXCLUDED_VIA_IMMEDIATE_HC_FAIL);

208

4

209

210

  // The threading here is complex. The cluster owns the only strong reference to the health

211

  // checker. It might go away when we post to the main thread from a worker thread. To deal with

212

  // this we use the following sequence of events:

213

  // 1) We capture a weak reference to the health checker and post it from worker thread to main

214

  //    thread.

215

  // 2) On the main thread, we make sure it is still valid (as the cluster may have been destroyed).

216

  // 3) Additionally, the host/session may also be gone by then so we check that also.

217

4

  std::weak_ptr<HealthCheckerImplBase> weak_this = shared_from_this();

218

4

  dispatcher_.post([weak_this, host]() -> void {

219

4

    std::shared_ptr<HealthCheckerImplBase> shared_this = weak_this.lock();

220

4

    if (shared_this == nullptr) {

221

1

      return;

222

1

223

224

3

    const auto session = shared_this->active_sessions_.find(host);

225

3

    if (session == shared_this->active_sessions_.end()) {

226

1

      return;

227

1

228

229

2

    session->second->setUnhealthy(envoy::data::core::v3::PASSIVE, /*retriable=*/false);

230

2

});

231

4

232

233

336

void HealthCheckerImplBase::start() {

234

338

  for (auto& host_set : cluster_.prioritySet().hostSetsPerPriority()) {

235

338

    addHosts(host_set->hosts());

236

338

237

336

238

239

HealthCheckerImplBase::ActiveHealthCheckSession::ActiveHealthCheckSession(

240

    HealthCheckerImplBase& parent, HostSharedPtr host)

241

353

    : host_(host), parent_(parent),

242

947

      interval_timer_(parent.dispatcher_.createTimer([this]() -> void { onIntervalBase(); })),

243

353

      timeout_timer_(parent.dispatcher_.createTimer([this]() -> void { onTimeoutBase(); })),

244

353

      time_source_(parent.dispatcher_.timeSource()) {

245

246

353

  if (!host->healthFlagGet(Host::HealthFlag::FAILED_ACTIVE_HC)) {

247

165

    parent.incHealthy();

248

165

249

250

353

  if (host->healthFlagGet(Host::HealthFlag::DEGRADED_ACTIVE_HC)) {

251

    parent.incDegraded();

252

253

353

254

255

353

HealthCheckerImplBase::ActiveHealthCheckSession::~ActiveHealthCheckSession() {

256

  // Make sure onDeferredDeleteBase() has been called. We should not reference our parent at this

257

  // point since we may have been deferred deleted.

258

353

  ASSERT(interval_timer_ == nullptr && timeout_timer_ == nullptr);

259

353

260

261

353

void HealthCheckerImplBase::ActiveHealthCheckSession::onDeferredDeleteBase() {

262

353

  HealthState state = HealthState::Unhealthy;

263

  // The session is about to be deferred deleted. Make sure all timers are gone and any

264

  // implementation specific state is destroyed.

265

353

  interval_timer_.reset();

266

353

  timeout_timer_.reset();

267

353

  if (!host_->healthFlagGet(Host::HealthFlag::FAILED_ACTIVE_HC)) {

268

161

    parent_.decHealthy();

269

161

    state = HealthState::Healthy;

270

161

271

353

  if (host_->healthFlagGet(Host::HealthFlag::DEGRADED_ACTIVE_HC)) {

272

    parent_.decDegraded();

273

274

353

  onDeferredDelete();

275

276

  // Run callbacks in case something is waiting for health checks to run which will now never run.

277

353

  if (first_check_) {

278

65

    parent_.runCallbacks(host_, HealthTransition::Unchanged, state);

279

65

280

353

281

282

869

void HealthCheckerImplBase::ActiveHealthCheckSession::handleSuccess(bool degraded) {

283

  // If we are healthy, reset the # of unhealthy to zero.

284

869

  num_unhealthy_ = 0;

285

286

869

  HealthTransition changed_state = HealthTransition::Unchanged;

287

288

869

  if (host_->healthFlagGet(Host::HealthFlag::FAILED_ACTIVE_HC)) {

289

    // If this is the first time we ever got a check result on this host, we immediately move

290

    // it to healthy. This makes startup faster with a small reduction in overall reliability

291

    // depending on the HC settings.

292

121

    if (first_check_ || ++num_healthy_ == parent_.healthy_threshold_) {

293

      // If the host moves to healthy, clear active HC timeout, which may be toggled off and on

294

      // while the host is unhealthy.

295

100

      host_->healthFlagClear(Host::HealthFlag::ACTIVE_HC_TIMEOUT);

296

      // A host that was told to exclude based on immediate failure, but is now passing, should

297

      // no longer be excluded.

298

100

      host_->healthFlagClear(Host::HealthFlag::EXCLUDED_VIA_IMMEDIATE_HC_FAIL);

299

100

      host_->healthFlagClear(Host::HealthFlag::FAILED_ACTIVE_HC);

300

100

      parent_.incHealthy();

301

100

      changed_state = HealthTransition::Changed;

302

100

      if (parent_.event_logger_) {

303

23

        parent_.event_logger_->logAddHealthy(parent_.healthCheckerType(), host_, first_check_);

304

23

305

101

    } else {

306

21

      changed_state = HealthTransition::ChangePending;

307

21

308

121

    host_->setLastHcPassTime(time_source_.monotonicTime());

309

121

310

311

869

  if (changed_state != HealthTransition::Changed && parent_.always_log_health_check_success_ &&

312

869

      parent_.event_logger_) {

313

1

    parent_.event_logger_->logSuccessfulHealthCheck(parent_.healthCheckerType(), host_);

314

1

315

316

869

  changed_state = clearPendingFlag(changed_state);

317

318

869

  if (degraded != host_->healthFlagGet(Host::HealthFlag::DEGRADED_ACTIVE_HC)) {

319

2

    if (degraded) {

320

1

      host_->healthFlagSet(Host::HealthFlag::DEGRADED_ACTIVE_HC);

321

1

      parent_.incDegraded();

322

1

      if (parent_.event_logger_) {

323

1

        parent_.event_logger_->logDegraded(parent_.healthCheckerType(), host_);

324

1

325

1

    } else {

326

1

      if (parent_.event_logger_) {

327

1

        parent_.event_logger_->logNoLongerDegraded(parent_.healthCheckerType(), host_);

328

1

329

1

      host_->healthFlagClear(Host::HealthFlag::DEGRADED_ACTIVE_HC);

330

1

331

332

    // This check ensures that we honor the decision made about Changed vs ChangePending in the

333

    // above block.

334

    // TODO(snowp): should there be degraded_threshold?

335

2

    if (changed_state == HealthTransition::Unchanged) {

336

2

      changed_state = HealthTransition::Changed;

337

2

338

2

339

340

869

  parent_.stats_.success_.inc();

341

869

  first_check_ = false;

342

869

  parent_.runCallbacks(host_, changed_state, HealthState::Healthy);

343

344

869

  timeout_timer_->disableTimer();

345

869

  interval_timer_->enableTimer(parent_.interval(HealthState::Healthy, changed_state));

346

869

347

348

namespace {

349

350

340

bool networkHealthCheckFailureType(envoy::data::core::v3::HealthCheckFailureType type) {

351

340

  return type == envoy::data::core::v3::NETWORK || type == envoy::data::core::v3::NETWORK_TIMEOUT;

352

340

353

354

} // namespace

355

356

HealthTransition HealthCheckerImplBase::ActiveHealthCheckSession::setUnhealthy(

357

215

    envoy::data::core::v3::HealthCheckFailureType type, bool retriable) {

358

  // If we are unhealthy, reset the # of healthy to zero.

359

215

  num_healthy_ = 0;

360

361

215

  HealthTransition changed_state = HealthTransition::Unchanged;

362

215

  if (!host_->healthFlagGet(Host::HealthFlag::FAILED_ACTIVE_HC)) {

363

125

    if ((!networkHealthCheckFailureType(type) && !retriable) ||

364

125

        ++num_unhealthy_ == parent_.unhealthy_threshold_) {

365

104

      host_->healthFlagSet(Host::HealthFlag::FAILED_ACTIVE_HC);

366

104

      parent_.decHealthy();

367

104

      changed_state = HealthTransition::Changed;

368

104

      if (parent_.event_logger_) {

369

75

        parent_.event_logger_->logEjectUnhealthy(parent_.healthCheckerType(), host_, type);

370

75

371

106

    } else {

372

21

      changed_state = HealthTransition::ChangePending;

373

21

374

125

375

376

  // In the case of network timeout and if the host is currently failed, set the timeout flag.

377

  // Otherwise clear it. This allows a host to toggle between timeout and failure if it's continuing

378

  // to fail for different reasons.

379

215

  if (type == envoy::data::core::v3::NETWORK_TIMEOUT &&

380

215

      host_->healthFlagGet(Host::HealthFlag::FAILED_ACTIVE_HC)) {

381

26

    host_->healthFlagSet(Host::HealthFlag::ACTIVE_HC_TIMEOUT);

382

189

  } else {

383

189

    host_->healthFlagClear(Host::HealthFlag::ACTIVE_HC_TIMEOUT);

384

189

385

386

215

  changed_state = clearPendingFlag(changed_state);

387

388

215

  if ((first_check_ || parent_.always_log_health_check_failures_) && parent_.event_logger_) {

389

70

    parent_.event_logger_->logUnhealthy(parent_.healthCheckerType(), host_, type, first_check_);

390

70

391

392

215

  parent_.stats_.failure_.inc();

393

215

  if (networkHealthCheckFailureType(type)) {

394

117

    parent_.stats_.network_failure_.inc();

395

130

  } else if (type == envoy::data::core::v3::PASSIVE) {

396

2

    parent_.stats_.passive_failure_.inc();

397

2

398

399

215

  first_check_ = false;

400

215

  parent_.runCallbacks(host_, changed_state, HealthState::Unhealthy);

401

215

  return changed_state;

402

215

403

404

void HealthCheckerImplBase::ActiveHealthCheckSession::handleFailure(

405

213

    envoy::data::core::v3::HealthCheckFailureType type, bool retriable) {

406

213

  HealthTransition changed_state = setUnhealthy(type, retriable);

407

  // It's possible that the previous call caused this session to be deferred deleted.

408

213

  if (timeout_timer_ != nullptr) {

409

204

    timeout_timer_->disableTimer();

410

204

411

412

213

  if (interval_timer_ != nullptr) {

413

204

    interval_timer_->enableTimer(parent_.interval(HealthState::Unhealthy, changed_state));

414

204

415

213

416

417

HealthTransition

418

1084

HealthCheckerImplBase::ActiveHealthCheckSession::clearPendingFlag(HealthTransition changed_state) {

419

1084

  if (host_->healthFlagGet(Host::HealthFlag::PENDING_ACTIVE_HC)) {

420

7

    host_->healthFlagClear(Host::HealthFlag::PENDING_ACTIVE_HC);

421

    // Even though the health value of the host might have not changed, we set this to Changed so

422

    // that the cluster can update its list of excluded hosts.

423

7

    return HealthTransition::Changed;

424

7

425

426

1077

  return changed_state;

427

1084

428

429

1160

void HealthCheckerImplBase::ActiveHealthCheckSession::onIntervalBase() {

430

1160

  onInterval();

431

1160

  timeout_timer_->enableTimer(parent_.timeout_);

432

1160

  parent_.stats_.attempt_.inc();

433

1160

434

435

36

void HealthCheckerImplBase::ActiveHealthCheckSession::onTimeoutBase() {

436

36

  onTimeout();

437

36

  handleFailure(envoy::data::core::v3::NETWORK_TIMEOUT);

438

36

439

440

352

void HealthCheckerImplBase::ActiveHealthCheckSession::onInitialInterval() {

441

352

  if (parent_.initial_jitter_.count() == 0) {

442

351

    onIntervalBase();

443

351

  } else {

444

1

    interval_timer_->enableTimer(

445

1

        std::chrono::milliseconds(parent_.intervalWithJitter(0, parent_.initial_jitter_)));

446

1

447

352

448

449

} // namespace Upstream

450

} // namespace Envoy