Job Already Delayed Check

Hey guys!
I have to start some recurrent jobs at the server start-up. To make sure a job doesn't delay twice I had to come up with some kind of a checkAlreadyDelayed function. It would be great to add a native kue check function or, even better, add a possibility to tell kue to make the job unique ( smth like kue.create('myjob').delay(1000).unique().save() ).
I could come up with a pull request at no time, just let me know if you find this case relevant.
So far I've been using this:

```
module.exports.checkJobAlreadyDelayed = function *(name) {
logger.debug('jobHelper.checkJobAlreadyDelayed -> init');
var ids = yield redis.zrangeAsync(kuePrefix + ':jobs:delayed', 0, -1);

var keys;
keys = _.map(ids, function (id) {
return kuePrefix + ':job:' + id.substr(id.indexOf('|')+1);
});

keys = yield _.map(keys, function (key) {
return redis.hgetAsync(key, 'type');
});

var resDelayed = _.reduce(keys, function (total, val) {
return total || (val === name);
}, false);

ids = yield redis.zrangeAsync(kuePrefix + ':jobs:inactive', 0, -1);

keys = _.map(ids, function (id) {
return kuePrefix + ':job:' + id.substr(id.indexOf('|')+1);
});

keys = yield _.map(keys, function (key) {
return redis.hgetAsync(key, 'type');
});

var resInactive = _.reduce(keys, function (total, val) {
return total || (val === name);
}, false);

return resDelayed || resInactive;
};
```

Start by reviewing Kue's job creation and delay flow, then inspect the Redis delayed and inactive job sets shown in the issue. Define how the requested duplicate check or unique-job behavior should work across those states; done means the chosen API reliably prevents or detects a job being delayed twice.

Written by the indexing model from the issue text.