* Last, but not least: consider a RateLimiter with rate of 1 permit per second, currently
   * completely unused, and an expensive acquire(100) request comes. It would be nonsensical to just
   * wait for 100 seconds, and /then/ start the actual task. Why wait without doing anything? A much
   * better approach is to /allow/ the request right away (as if it was an acquire(1) request

participant dep as Dep with yield
participant operation as Path Operation
participant tasks as Background tasks

    Note over client,operation: Can raise exceptions, including HTTPException
    client ->> dep: Start request
    Note over dep: Run code up to yield
    opt raise Exception
        dep -->> handler: Raise Exception
        handler -->> client: HTTP error response
    end

	bucketMetas := make([]BucketMetadata, len(buckets))
	for index := range buckets {
		index := index
		g.Go(func() error {
			// Sleep and stagger to avoid blocked CPU and thundering
			// herd upon start up sequence.
			time.Sleep(25*time.Millisecond + time.Duration(rand.Int63n(int64(100*time.Millisecond))))

			_, _ = sys.objAPI.HealBucket(ctx, buckets[index], madmin.HealOpts{Recreate: true})

type StorageInfo = madmin.StorageInfo

// objectHistogramInterval is an interval that will be
// used to report the histogram of objects data sizes
type objectHistogramInterval struct {
	name       string
	start, end int64
}

const (
	// dataUsageBucketLenV1 must be length of ObjectsHistogramIntervalsV1
	dataUsageBucketLenV1 = 7
	// dataUsageBucketLen must be length of ObjectsHistogramIntervals
	dataUsageBucketLen  = 11

This guide will help you to...

* maximize the chance of your changes being accepted
* work on the Gradle code base
* get help if you encounter trouble

## Before you start

Before starting to work on a feature or a bug fix, please open an issue to discuss the use case or bug with us. This can save everyone a lot of time and frustration.

// Some TF ops need a step container to be set to limit the lifetime of some
// resources (mostly TensorArray and Stack, used in while loop gradients in
// graph mode). Calling this on a context tells it to start a step.
TF_CAPI_EXPORT extern void TFE_ContextStartStep(TFE_Context* ctx);

// Ends a step. When there is no active step (that is, every started step has

      TF_SetStatus(status, TF_GetCode(status_.get()),
                   TF_Message(status_.get()));
      // Reset the member `status_` so future op executions (after recovery from
      // the bad `status`) start with an OK status.
      TF_SetStatus(status_.get(), TF_OK, "");
    }
    cancellation_manager_ = nullptr;
    execution_state_ = ExecutionState::kIdle;
    result = std::move(op_outputs_);
  }

					podPort = v.TargetPort.IntValue()
					break
				}
			}
			f, err := e.client.NewPortForwarder(pod.Name, pod.Namespace, "", 0, podPort)
			if err != nil {
				return nil, err
			}
			if err := f.Start(); err != nil {
				return nil, err
			}
			address = fmt.Sprintf("https://%s%s", f.Address(), *cc.Service.Path)
			defer func() {
				f.Close()
				f.WaitForStop()
			}()
		}

adjust the `max_sleep` and `max_io` values thereby increasing the healing speed. The delays between each operation of the healer can be adjusted by the `mc admin config set alias/ heal max_sleep=1s` and maximum concurrent requests allowed before we start slowing things down can be configured with `mc admin config set alias/ heal max_io=30` . By default the wait delay is `250ms` beyond 100 concurrent operations. This means the healer will sleep *250 milliseconds* at max for each heal operation if there...

	default:
		return nil, errInvalidASTNode
	}
}

func (e *Between) evalBetweenNode(r Record, arg *Value, tableAlias string) (*Value, error) {
	stVal, stErr := e.Start.evalNode(r, tableAlias)
	if stErr != nil {
		return nil, stErr
	}

	endVal, endErr := e.End.evalNode(r, tableAlias)
	if endErr != nil {
		return nil, endErr
	}