-
Notifications
You must be signed in to change notification settings - Fork 1
/
driver.go
401 lines (320 loc) · 10.4 KB
/
driver.go
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
package memoryjobs
import (
"context"
"strconv"
"sync"
"sync/atomic"
"time"
"github.com/roadrunner-server/api/v4/plugins/v3/jobs"
"github.com/roadrunner-server/errors"
jprop "go.opentelemetry.io/contrib/propagators/jaeger"
"go.opentelemetry.io/otel"
"go.opentelemetry.io/otel/propagation"
sdktrace "go.opentelemetry.io/otel/sdk/trace"
"go.opentelemetry.io/otel/trace"
"go.uber.org/zap"
)
const (
pluginName string = "memory"
prefetch string = "prefetch"
goroutinesMax uint64 = 1000
// v2023.1 OTEL
tracerName string = "jobs"
)
var _ jobs.Driver = (*Driver)(nil)
type Configurer interface {
// UnmarshalKey takes a single key and unmarshal it into a Struct.
UnmarshalKey(name string, out any) error
// Has checks if config section exists.
Has(name string) bool
}
type Config struct {
Priority int64 `mapstructure:"priority"`
Prefetch int64 `mapstructure:"prefetch"`
}
type Driver struct {
cfg *Config
// delayed messages
delayed *int64
// currently processing
msgInFlight *int64
// prefetch
msgInFlightLimit *int64
cond sync.Cond
tracer *sdktrace.TracerProvider
log *zap.Logger
pipeline atomic.Pointer[jobs.Pipeline]
pq jobs.Queue
localQueue chan *Item
prop propagation.TextMapPropagator
// time.sleep goroutines max number
goroutines uint64
priority int64
listeners uint32
stopped uint64
stopCh chan struct{}
}
// FromConfig initializes kafka pipeline from the configuration
func FromConfig(
tracer *sdktrace.TracerProvider,
configKey string,
log *zap.Logger,
cfg Configurer,
pipeline jobs.Pipeline,
pq jobs.Queue) (*Driver, error) {
const op = errors.Op("new_in_memory_pipeline")
if tracer == nil {
tracer = sdktrace.NewTracerProvider()
}
jb := &Driver{
stopped: 0,
tracer: tracer,
cond: sync.Cond{L: &sync.Mutex{}},
log: log,
pq: pq,
goroutines: 0,
msgInFlight: toPtr(int64(0)),
delayed: toPtr(int64(0)),
stopCh: make(chan struct{}),
}
err := cfg.UnmarshalKey(configKey, &jb.cfg)
if err != nil {
return nil, errors.E(op, err)
}
if jb.cfg == nil {
return nil, errors.E(op, errors.Errorf("config not found by provided key: %s", configKey))
}
if jb.cfg.Prefetch == 0 {
jb.cfg.Prefetch = 100_000
}
jb.msgInFlightLimit = toPtr(jb.cfg.Prefetch)
if jb.cfg.Priority == 0 {
jb.cfg.Priority = 10
}
jb.priority = jb.cfg.Priority
jb.pipeline.Store(&pipeline)
// initialize a local queue
jb.localQueue = make(chan *Item, 100_000)
jb.prop = propagation.NewCompositeTextMapPropagator(propagation.TraceContext{}, propagation.Baggage{}, jprop.Jaeger{})
otel.SetTextMapPropagator(jb.prop)
return jb, nil
}
// FromPipeline initializes pipeline on-the-fly
func FromPipeline(
tracer *sdktrace.TracerProvider,
pipeline jobs.Pipeline,
log *zap.Logger,
pq jobs.Queue,
) (*Driver, error) {
pref, err := strconv.ParseInt(pipeline.String(prefetch, "100000"), 10, 64)
if err != nil {
return nil, err
}
if tracer == nil {
tracer = sdktrace.NewTracerProvider()
}
dr := &Driver{
stopped: 0,
tracer: tracer,
log: log,
pq: pq,
cond: sync.Cond{L: &sync.Mutex{}},
localQueue: make(chan *Item, 100_000),
goroutines: 0,
msgInFlight: toPtr(int64(0)),
msgInFlightLimit: toPtr(pref),
delayed: toPtr(int64(0)),
priority: pipeline.Priority(),
stopCh: make(chan struct{}),
}
dr.pipeline.Store(&pipeline)
dr.prop = propagation.NewCompositeTextMapPropagator(propagation.TraceContext{}, propagation.Baggage{}, jprop.Jaeger{})
otel.SetTextMapPropagator(dr.prop)
return dr, nil
}
func (c *Driver) Push(ctx context.Context, jb jobs.Message) error {
const op = errors.Op("in_memory_push")
ctx, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_push")
defer span.End()
// check if the pipeline registered
pipe := *c.pipeline.Load()
if pipe == nil {
return errors.E(op, errors.Errorf("no such pipeline: %s", jb.GroupID()))
}
err := c.handleItem(ctx, fromJob(jb))
if err != nil {
return errors.E(op, err)
}
return nil
}
func (c *Driver) State(ctx context.Context) (*jobs.State, error) {
_, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_state")
defer span.End()
pipe := *c.pipeline.Load()
return &jobs.State{
Pipeline: pipe.Name(),
Priority: uint64(pipe.Priority()),
Driver: pipe.Driver(),
Queue: pipe.Name(),
Active: atomic.LoadInt64(c.msgInFlight),
Delayed: atomic.LoadInt64(c.delayed),
Ready: ready(atomic.LoadUint32(&c.listeners)),
}, nil
}
func (c *Driver) Run(ctx context.Context, pipe jobs.Pipeline) error {
const op = errors.Op("in_memory_jobs_run")
_, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_run")
defer span.End()
t := time.Now().UTC()
l := atomic.LoadUint32(&c.listeners)
// listener already active
if l == 1 {
c.log.Warn("listener already in the active state")
return errors.E(op, errors.Str("listener already in the active state"))
}
c.consume()
atomic.StoreUint32(&c.listeners, 1)
c.log.Debug("pipeline was started", zap.String("driver", pipe.Driver()), zap.String("pipeline", pipe.Name()), zap.String("start", time.Now().UTC().String()), zap.String("elapsed", time.Since(t).String()))
return nil
}
func (c *Driver) Pause(ctx context.Context, p string) error {
start := time.Now().UTC()
_, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_pause")
defer span.End()
pipe := *c.pipeline.Load()
if pipe.Name() != p {
return errors.Errorf("no such pipeline: %s", pipe.Name())
}
l := atomic.LoadUint32(&c.listeners)
// no active listeners
if l == 0 {
return errors.Str("no active listeners, nothing to pause")
}
atomic.AddUint32(&c.listeners, ^uint32(0))
// stop the Driver
c.stopCh <- struct{}{}
c.log.Debug("pipeline was paused", zap.String("driver", pipe.Driver()), zap.String("pipeline", pipe.Name()), zap.String("start", time.Now().UTC().String()), zap.String("elapsed", time.Since(start).String()))
return nil
}
func (c *Driver) Resume(ctx context.Context, p string) error {
start := time.Now().UTC()
_, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_resume")
defer span.End()
pipe := *c.pipeline.Load()
if pipe.Name() != p {
return errors.Errorf("no such pipeline: %s", pipe.Name())
}
l := atomic.LoadUint32(&c.listeners)
// listener already active
if l == 1 {
return errors.Str("memory listener is already in the active state")
}
// resume the Driver on the same channel
c.consume()
atomic.StoreUint32(&c.listeners, 1)
c.log.Debug("pipeline was resumed", zap.String("driver", pipe.Driver()), zap.String("pipeline", pipe.Name()), zap.String("start", time.Now().UTC().String()), zap.String("elapsed", time.Since(start).String()))
return nil
}
func (c *Driver) Stop(ctx context.Context) error {
start := time.Now().UTC()
_, span := trace.SpanFromContext(ctx).TracerProvider().Tracer(tracerName).Start(ctx, "in_memory_stop")
defer span.End()
pipe := *c.pipeline.Load()
_ = c.pq.Remove(pipe.Name())
select {
case c.stopCh <- struct{}{}:
default:
break
}
// just to be sure, that queue won't block
c.cond.Signal()
// close localQueue channel
close(c.localQueue)
// help GC
c.localQueue = nil
atomic.StoreUint64(&c.stopped, 1)
c.log.Debug("pipeline was stopped", zap.String("driver", pipe.Driver()), zap.String("pipeline", pipe.Name()), zap.String("start", time.Now().UTC().String()), zap.String("elapsed", time.Since(start).String()))
return nil
}
func (c *Driver) handleItem(ctx context.Context, msg *Item) error {
const op = errors.Op("in_memory_handle_request")
c.prop.Inject(ctx, propagation.HeaderCarrier(msg.headers))
// handle timeouts
// theoretically, some bad user may send millions requests with a delay and produce a billion (for example)
// goroutines here. We should limit goroutines here.
if msg.Options.Delay > 0 {
// if we have 1000 goroutines waiting on the delay - reject 1001
if atomic.LoadUint64(&c.goroutines) >= goroutinesMax {
return errors.E(op, errors.Str("max concurrency number reached"))
}
go func(jj *Item) {
atomic.AddUint64(&c.goroutines, 1)
atomic.AddInt64(c.delayed, 1)
defer atomic.AddUint64(&c.goroutines, ^uint64(0))
time.Sleep(jj.Options.DelayDuration())
select {
case c.localQueue <- jj:
default:
c.log.Warn("can't push job", zap.String("error", "local queue closed or full"))
}
}(msg)
return nil
}
// insert to the local, limited pipeline
select {
case c.localQueue <- msg:
return nil
case <-ctx.Done():
return errors.E(op, errors.Errorf("local pipeline is full, consider to increase prefetch number, current limit: %d, context error: %v", c.cfg.Prefetch, ctx.Err()))
}
}
func (c *Driver) consume() {
go func() {
// redirect
for {
select {
case item, ok := <-c.localQueue:
if !ok {
c.log.Debug("ephemeral local queue closed")
return
}
ctx := otel.GetTextMapPropagator().Extract(context.Background(), propagation.HeaderCarrier(item.headers))
ctx, span := c.tracer.Tracer(tracerName).Start(ctx, "in_memory_listener")
c.cond.L.Lock()
for atomic.LoadInt64(c.msgInFlight) >= atomic.LoadInt64(c.msgInFlightLimit) {
c.log.Debug("prefetch limit was reached, waiting for the jobs to be processed", zap.Int64("current", atomic.LoadInt64(c.msgInFlight)), zap.Int64("limit", atomic.LoadInt64(c.msgInFlightLimit)))
// wait for the jobs to be processed
c.cond.Wait()
}
if item.Priority() == 0 {
item.Options.Priority = c.priority
}
// set requeue channel
item.Options.requeueFn = c.handleItem
item.Options.msgInFlight = c.msgInFlight
item.Options.delayed = c.delayed
item.Options.cond = &c.cond
item.Options.stopped = &c.stopped
if item.headers == nil {
item.headers = make(map[string][]string, 1)
}
// inject OTEL headers
c.prop.Inject(ctx, propagation.HeaderCarrier(item.headers))
c.pq.Insert(item)
// increase number of the active jobs
atomic.AddInt64(c.msgInFlight, 1)
c.log.Debug("message pushed to the priority queue", zap.Int64("current", atomic.LoadInt64(c.msgInFlight)), zap.Int64("limit", atomic.LoadInt64(c.msgInFlightLimit)))
c.cond.L.Unlock()
span.End()
case <-c.stopCh:
return
}
}
}()
}
func ready(r uint32) bool {
return r > 0
}
func toPtr[T any](v T) *T {
return &v
}