Fixed critical bug that didn't allow ingesting to different clusters

Fixes

• Critical bug - When ingesting to multiple clusters all data is sent to one cluster.
  As always, we recommend re-using clients and ingestors whenever possible.

Lower memory consumption and control, Fixed and added formats, Security fixes

Breaking Changes

• The minimal supported go version has been bumped from 1.13 to 1.16.
  1.13 has been EOL for a while, and 1.16 is required for secure versions of our dependencies.

Features

• Lower memory consumption when ingesting from file - instead of using a huge static buffer to upload files to an azure storage blob, the default is now using a synchronized buffer pool. This can reduce memory consumption by a factor of 10x.
• An WithStaticBuffer option to ingesting is available when the old behavior is requested, with the added control of the buffers count and size
• Added support for the singlejson format

Fixes

• Fixed security alert for github.com/dgrijalva/jwt-go by updating inner dependencies to avoid the package
• Fixed security alert for github.com/satori/go.uuid by updating inner dependencies to avoid the package
• Fixed typos that caused w3clog and sstream files to have the wrong extension

Fixed management query bugs

• Management queries now work correctly with complex queries or multiple tables. #55

Fixed synapse integration

• The fix for synapse URLs now works in all flows

Added CreationTime property

New features:

• Added additional ingestion property CreationTime, which allows to specify the creation time of ingested data in kusto's extents. When unspecifies default to now()

Added support for synapse clusters

• Added special support for synapse by fixing audience for URIs having azuresynapse like segments

Added auth token caching, Fixed e2e tests

Fixes:

• Ingestion does not call the cluster for identity token for each file/reader
• E2E tests now run properly

Documentation for Ingestion Status Reporting (Breaking)

Modify documentation to reflect changes in ingestion APIs

Add Support for Ingestion Status Reporting (Breaking)

Table Based Ingestion Status Reporting - Breaking Change

You can use Kusto Go SDK to get table-based status reporting of ingestion operations.
Ingestion commands now return an error and a channel that can be waited upon for a final status.
If the error is not nil, the operation has failed locally.
If the error is nil and Table Status Reporting option was used, the SDK user can wait on the channel for a success (nil) or failure (Error) status.

Note!
This feature is not suitable for users running ingestion at high rates, and may slow down the ingestion operation.

Usage:

// Upload a file with status reporting.
status, err := ingestor.FromFile(ctx, "/path/to/file", ingest.ReportResultToTable())
if err != nil {
	// The ingestion command failed to be sent, Do something
}

err = <-status.Wait(ctx)
if err != nil {
	// the operation complete with an error
	if ingest.IsRetryable(err) {
		// Handle retries
	} else {
		// inspect the failure
		// statusCode, _ := ingest.GetIngestionStatus(err)
		// failureStatus, _ := ingest.GetIngestionFailureStatus(err)
	}
}

Fix Timespan marshal, Remove "show ingestion mappings" call

• Fixes Timepsan to marshal correctly (had nanosecond precision, Timespan has "tick" precision)
• Removes "show ingestion mapping" calls that validated mapping references exist. This was causing cluster issues because our last cache time wasn't being written. However, we are moving in the next release to ingestion status calls for debugging, which are significantly better than this.