Documentation ¶
Overview ¶
Package rate The key observation and some code is borrowed from golang.org/x/time/rate/rate.go
Index ¶
- type BurstLimiter
- func (lim *BurstLimiter) Allow() bool
- func (lim *BurstLimiter) AllowN(n int) bool
- func (lim *BurstLimiter) Burst() int
- func (lim *BurstLimiter) GetToken() (ok bool)
- func (lim *BurstLimiter) GetTokenN(n int) (ok bool)
- func (lim *BurstLimiter) PutToken()
- func (lim *BurstLimiter) PutTokenN(n int)
- func (lim *BurstLimiter) Reserve(ctx context.Context) *Reservation
- func (lim *BurstLimiter) ReserveN(ctx context.Context, n int) *Reservation
- func (lim *BurstLimiter) SetBurst(newBurst int)
- func (lim *BurstLimiter) Tokens() int
- func (lim *BurstLimiter) Wait(ctx context.Context) (err error)
- func (lim *BurstLimiter) WaitN(ctx context.Context, n int) (err error)
- type Reservation
Examples ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type BurstLimiter ¶
type BurstLimiter struct {
// contains filtered or unexported fields
}
BurstLimiter Informally, in any large enough time interval, the BurstLimiter limits the burst tokens, with a maximum burst size of b events. As a special case, if r == Inf (the infinite rate), b is ignored. See https://en.wikipedia.org/wiki/Token_bucket for more about token buckets.
Reorder Buffer It allows instructions to be committed in-order. - Allocated by `Reserve` or `ReserveN` into account when allowing future events - Wait by `Wait` or `WaitN` blocks until lim permits n events to happen. - Allow and Wait Complete by `PutToken` or `PutTokenN` - Reserve Complete by `Cancel` of the Reservation self, GC Cancel supported See https://en.wikipedia.org/wiki/Re-order_buffer for more about Reorder buffer. See https://web.archive.org/web/20040724215416/http://lgjohn.okstate.edu/6253/lectures/reorder.pdf for more about Reorder buffer.
The zero value is a valid BurstLimiter, but it will reject all events. Use NewFullBurstLimiter to create non-zero Limiters.
BurstLimiter has three main methods, Allow, Reserve, and Wait. Most callers should use Wait for token bucket. Most callers should use Reserve for Reorder buffer.
Each of the three methods consumes a single token. They differ in their behavior when no token is available. If no token is available, Allow returns false. If no token is available, Reserve returns a reservation for a future token and the amount of time the caller must wait before using it. If no token is available, Wait blocks until one can be obtained or its associated context.Context is canceled.
The methods AllowN, ReserveN, and WaitN consume n tokens.
func NewEmptyBurstLimiter ¶
func NewEmptyBurstLimiter(b int) *BurstLimiter
NewEmptyBurstLimiter returns a new BurstLimiter inited with zero tokens that allows events up to burst b and permits bursts of at most b tokens.
Example ¶
package main import ( "context" "fmt" "sync" "time" "github.com/searKing/golang/go/time/rate" ) func main() { const ( burst = 3 concurrency = 2 ) limiter := rate.NewEmptyBurstLimiter(burst) ctx, cancel := context.WithTimeout(context.Background(), time.Hour) defer cancel() fmt.Printf("tokens left: %d\n", limiter.Tokens()) // expect not allowed, as limiter is initialized with empty tokens(0) if limiter.Allow() { fmt.Printf("allow passed\n") } else { fmt.Printf("allow refused\n") } // fill one token limiter.PutToken() fmt.Printf("tokens left: %d\n", limiter.Tokens()) // expect allowed, as limiter is filled with one token(1) if limiter.Allow() { fmt.Printf("allow passed\n") } else { fmt.Printf("allow refused\n") } fmt.Printf("tokens left: %d\n", limiter.Tokens()) var mu sync.Mutex var wg sync.WaitGroup for i := 0; i < concurrency; i++ { wg.Add(1) go func() { defer wg.Done() //fmt.Printf("%03d %s\n", i, time.Now().Format(time.RFC3339)) mu.Lock() fmt.Printf("Wait 1 Token, tokens left: %d\n", limiter.Tokens()) mu.Unlock() err := limiter.Wait(ctx) if err != nil { mu.Lock() fmt.Printf("err: %s\n", err.Error()) mu.Unlock() return } mu.Lock() fmt.Printf("Got 1 Token, tokens left: %d\n", limiter.Tokens()) mu.Unlock() }() } time.Sleep(10 * time.Millisecond) for i := 0; i < concurrency; i++ { time.Sleep(10 * time.Millisecond) mu.Lock() fmt.Printf("PutToken #%d: before tokens left: %d\n", i, limiter.Tokens()) // fill one token limiter.PutToken() fmt.Printf("PutToken #%d: after tokens left: %d\n", i, limiter.Tokens()) mu.Unlock() } wg.Wait() fmt.Printf("tokens left: %d\n", limiter.Tokens()) // expect allowed, as limiter is filled with one token(1) if limiter.Allow() { fmt.Printf("allow passed\n") } else { fmt.Printf("allow refused\n") } fmt.Printf("tokens left: %d\n", limiter.Tokens()) // expect not allowed, as limiter is initialized with empty tokens(0) if limiter.Allow() { fmt.Printf("allow passed\n") } else { fmt.Printf("allow refused\n") } }
Output: tokens left: 0 allow refused tokens left: 1 allow passed tokens left: 0 Wait 1 Token, tokens left: 0 Wait 1 Token, tokens left: 0 PutToken #0: before tokens left: 0 PutToken #0: after tokens left: 0 Got 1 Token, tokens left: 0 PutToken #1: before tokens left: 0 PutToken #1: after tokens left: 0 Got 1 Token, tokens left: 0 tokens left: 0 allow refused tokens left: 0 allow refused
func NewFullBurstLimiter ¶
func NewFullBurstLimiter(b int) *BurstLimiter
NewFullBurstLimiter returns a new BurstLimiter inited with full tokens that allows events up to burst b and permits bursts of at most b tokens.
Example ¶
package main import ( "context" "fmt" "runtime" "time" "github.com/searKing/golang/go/time/rate" ) func main() { const ( burst = 3 ) limiter := rate.NewFullBurstLimiter(burst) ctx, cancel := context.WithTimeout(context.Background(), time.Second) defer cancel() // expect dropped, as limiter is initialized with full tokens(3) limiter.PutToken() for i := 0; ; i++ { //fmt.Printf("%03d %s\n", i, time.Now().Format(time.RFC3339)) fmt.Printf("Wait %03d, tokens left: %d\n", i, limiter.Tokens()) err := limiter.Wait(ctx) if err != nil { fmt.Printf("err: %s\n", err.Error()) return } fmt.Printf("Got %03d, tokens left: %d\n", i, limiter.Tokens()) // actor mocked by gc runtime.GC() if i == 0 { // refill one token limiter.PutToken() } } }
Output: Wait 000, tokens left: 3 Got 000, tokens left: 2 Wait 001, tokens left: 3 Got 001, tokens left: 2 Wait 002, tokens left: 2 Got 002, tokens left: 1 Wait 003, tokens left: 1 Got 003, tokens left: 0 Wait 004, tokens left: 0 err: context deadline exceeded
func (*BurstLimiter) Allow ¶
func (lim *BurstLimiter) Allow() bool
Allow is shorthand for AllowN(time.Now(), 1). 当没有可用或足够的事件时,返回false
func (*BurstLimiter) AllowN ¶
func (lim *BurstLimiter) AllowN(n int) bool
AllowN reports whether n events may happen at time now. AllowN is shorthand for GetTokenN. Use this method if you intend to drop / skip events that exceed the rate limit. Otherwise, use Reserve or Wait. 当没有可用或足够的事件时,返回false
func (*BurstLimiter) Burst ¶
func (lim *BurstLimiter) Burst() int
Burst returns the maximum burst size. Burst is the maximum number of tokens that can be consumed in a single call to Allow, Reserve, or Wait, so higher Burst values allow more events to happen at once. A zero Burst allows no events, unless limit == Inf.
func (*BurstLimiter) GetToken ¶
func (lim *BurstLimiter) GetToken() (ok bool)
GetToken is shorthand for GetTokenN(ctx, 1).
func (*BurstLimiter) GetTokenN ¶
func (lim *BurstLimiter) GetTokenN(n int) (ok bool)
GetTokenN returns true if token is got
func (*BurstLimiter) PutToken ¶
func (lim *BurstLimiter) PutToken()
PutToken is shorthand for PutTokenN(ctx, 1).
func (*BurstLimiter) PutTokenN ¶
func (lim *BurstLimiter) PutTokenN(n int)
func (*BurstLimiter) Reserve ¶
func (lim *BurstLimiter) Reserve(ctx context.Context) *Reservation
Reserve is shorthand for ReserveN(1). 当没有可用或足够的事件时,返回 Reservation,和要等待多久才能获得足够的事件。
Example ¶
package main import ( "context" "fmt" "math/rand" "sync" "time" "github.com/searKing/golang/go/time/rate" ) func main() { const ( burst = 1 n = 10 ) limiter := rate.NewFullBurstLimiter(burst) ctx, cancel := context.WithTimeout(context.Background(), time.Hour) defer cancel() // expect dropped, as limiter is initialized with full tokens(3) limiter.PutToken() type Reservation struct { index int r *rate.Reservation } var mu sync.Mutex var wg sync.WaitGroup var rs []*Reservation for i := 0; i < n; i++ { //fmt.Printf("%03d %s\n", i, time.Now().Format(time.RFC3339)) fmt.Printf("Reserve %03d\n", i) r := &Reservation{ index: i, r: limiter.Reserve(ctx), } if i%2 == rand.Intn(2)%2 { rs = append(rs, r) continue } wg.Add(1) go func() { defer wg.Done() //fmt.Printf("%03d %s\n", r.index, time.Now().Format(time.RFC3339)) //fmt.Printf("Wait %03d\n", r.index) err := r.r.Wait(ctx) if err != nil { mu.Lock() fmt.Printf("err: %s\n", err.Error()) mu.Unlock() } mu.Lock() fmt.Printf("%03d Got 1 Token, tokens left: %d\n", r.index, limiter.Tokens()) mu.Unlock() r.r.PutToken() }() } for i := 0; i < len(rs); i++ { r := rs[i] wg.Add(1) go func() { defer wg.Done() //fmt.Printf("%03d %s\n", r.index, time.Now().Format(time.RFC3339)) //fmt.Printf("Wait %03d\n", r.index) err := r.r.Wait(ctx) if err != nil { mu.Lock() fmt.Printf("err: %s\n", err.Error()) mu.Unlock() } mu.Lock() fmt.Printf("%03d Got 1 Token, tokens left: %d\n", r.index, limiter.Tokens()) mu.Unlock() r.r.PutToken() }() } wg.Wait() }
Output: Reserve 000 Reserve 001 Reserve 002 Reserve 003 Reserve 004 Reserve 005 Reserve 006 Reserve 007 Reserve 008 Reserve 009 000 Got 1 Token, tokens left: 0 001 Got 1 Token, tokens left: 0 002 Got 1 Token, tokens left: 0 003 Got 1 Token, tokens left: 0 004 Got 1 Token, tokens left: 0 005 Got 1 Token, tokens left: 0 006 Got 1 Token, tokens left: 0 007 Got 1 Token, tokens left: 0 008 Got 1 Token, tokens left: 0 009 Got 1 Token, tokens left: 0
func (*BurstLimiter) ReserveN ¶
func (lim *BurstLimiter) ReserveN(ctx context.Context, n int) *Reservation
ReserveN returns a Reservation that indicates how long the caller must wait before n events happen. The BurstLimiter takes this Reservation into account when allowing future events. ReserveN returns false if n exceeds the BurstLimiter's burst size. Usage example:
// Allocate: The dispatch stage reserves space in the reorder buffer for instructions in program order. r := lim.ReserveN(context.Background(), 1) if !r.OK() { // Not allowed to act! Did you remember to set lim.burst to be > 0 ? return } // Execute: out-of-order execution Act() // Wait: The complete stage must wait for instructions to finish execution. if err:= r.Wait(); err!=nil { // Not allowed to act! Reservation or context canceled ? return } // Complete: Finished instructions are allowed to write results in order into the architected registers. // It allows instructions to be committed in-order. defer r.PutToken() // Execute: in-order execution Act()
Use this method if you wish to wait and slow down in accordance with the rate limit without dropping events. If you need to respect a deadline or cancel the delay, use Wait instead. To drop or skip events exceeding rate limit, use Allow instead. 当没有可用或足够的事件时,返回 Reservation,和要等待多久才能获得足够的事件。 See https://en.wikipedia.org/wiki/Re-order_buffer for more about Reorder buffer. See https://web.archive.org/web/20040724215416/http://lgjohn.okstate.edu/6253/lectures/reorder.pdf for more about Reorder buffer.
func (*BurstLimiter) SetBurst ¶
func (lim *BurstLimiter) SetBurst(newBurst int)
SetBurst sets a new burst size for the limiter.
func (*BurstLimiter) Tokens ¶
func (lim *BurstLimiter) Tokens() int
Tokens returns the token nums unconsumed.
func (*BurstLimiter) Wait ¶
func (lim *BurstLimiter) Wait(ctx context.Context) (err error)
Wait is shorthand for WaitN(ctx, 1).
func (*BurstLimiter) WaitN ¶
func (lim *BurstLimiter) WaitN(ctx context.Context, n int) (err error)
WaitN blocks until lim permits n events to happen. It returns an error if n exceeds the BurstLimiter's burst size, the Context is canceled, or the expected wait time exceeds the Context's Deadline. The burst limit is ignored if the rate limit is Inf.
type Reservation ¶
type Reservation struct {
// contains filtered or unexported fields
}
A Reservation holds information about events that are permitted by a BurstLimiter to happen after a delay. A Reservation may be canceled, which may enable the BurstLimiter to permit additional events.
func (*Reservation) Cancel ¶
func (r *Reservation) Cancel()
Cancel indicates that the reservation holder will not perform the reserved action and reverses the effects of this Reservation on the rate limit as much as possible, considering that other reservations may have already been made.
func (*Reservation) OK ¶
func (r *Reservation) OK() bool
OK returns whether the limiter can provide the requested number of tokens within the maximum wait time. If OK is false, Delay returns InfDuration, and Cancel does nothing.
func (*Reservation) PutToken ¶ added in v1.2.56
func (r *Reservation) PutToken()
PutToken (as Complete): refill all tokens taken by the Reservation back to BurstLimiter. PutToken is shorthand for Cancel().
func (*Reservation) Ready ¶ added in v1.2.56
func (r *Reservation) Ready() bool
Ready returns whether the limiter can provide the requested number of tokens within the maximum wait time. If Ready is false, Wait returns nil directly, and Cancel or GC does put back the token reserved in the Reservation. If Ready is false, WaitN blocks until lim permits n events to happen.