Fixed saving unique Buffer types

v2.3.4

Fixed saving unique Buffer types

Added new options to break JSON object/array into multiple tags

The new index option can be set for any JSON object or array tag (all except id tag).

When set to true on a tag, it will break all data inside the value to multiple tags for easy search in Jaeger UI, regardless of the backend storage used for Jaeger and its indexing capabilities.

Example:

data: {
  account: { 
    id: 1
  },
  users: [
    {
      name: 'John' 
    }
  ]
}

Will be set to the following tags:

'data.account.id': 1
'data.users[0].name': 'John' 

Now you'll be able to search in Jaeger UI for specific inner properties:

data.account.id=1 data.users[0].name=John

Check the docs for config example.

Added support for new option `hideErrors`

This option will prevent tagging of selected services errors with error=false and their sampling priority won't be set to 1.

When the option is specified with a service path and an array of error codes, it will "hide" these errors when filtering Jaeger UI with error=true. This way, a user can exclude certain errors that are part of his normal implementation flow and have a much more clear focus on the real errors in his system.

For examples, check the docs

v2.2.3

Updated docs and bumped version to 2.2.3

v2.2.2

Bumped version to 2.2.2

v2.2.1

Bumped version to 2.2.1

v2.2.0

Updated dependencies and bumped version to 2.2.0

`opentracingError` hook now expects options object to handle early errors

Change opentracingError() to opentracingError(config.opentracing.options).

Sometimes, opentracingError hook can be called when an error is thrown before calling the opentracingBegin hook. In that case, opentracingError will now call opentracingBegin to create the required span before trying to use it.

Added support for `feathers-distributed`

feathers-distributed

2.1.1: * Added support for `feathers-distribution`

* Updated docs with Mac OS X issue workaround
* Bumped version to 2.1.1