I use a pattern similar to one already posted, but with the following difference:

public sealed class Logging
{
    static Logging instance = null;
    static readonly object lockObj = new object();

    private Logging()
    {
    }

    public static Logging Logger
    {
        get
        {
            **if (instance == null)**
            {
                 lock (lockObj)
                 {
                     if (instance == null)
                     {
                         instance = new Logging();
                     }
                    return instance;
                 }
            }
        }
    }

}

The reason being that it avoids calling lock every time, which can help you with performance. So you check for null, then, if it is null, lock it. Then you have to check for null again (because someone may have come in right that second) but then you should be ok. That way, you'll only hit the lock the first time (or two), and blow past it without locking the rest of the time.

I use a pattern similar to one already posted, but with the following difference:

public sealed class Logging
{
    static Logging instance = null;
    static readonly object lockObj = new object();

    private Logging()
    {
    }

    public static Logging Logger
    {
        get
        {
            **if (instance == null)**
            {
                 lock (lockObj)
                 {
                     if (instance == null)
                     {
                         instance = new Logging();
                     }
                    
                 }
            }
            return instance;
        }
    }

}

The reason being that it avoids calling lock every time, which can help you with performance. So you check for null, then, if it is null, lock it. Then you have to check for null again (because someone may have come in right that second) but then you should be ok. That way, you'll only hit the lock the first time (or two), and blow past it without locking the rest of the time.