Question

I have just started looking at ss and redis. i am using microsoft redis implementation. with compression turned on, the dump.rdb is growing too fast.

I would like to save per second process stats. example object.

public class PerfData
{
    public long Id { get; set; }    
    public DateTime TimeStamp { get; set; }
    public string ProcessName { get; set; }
    public int ProcessId { get; set; }
    public TimeSpan TotalProcessorTime { get; set; }
    public TimeSpan UserProcessorTime { get; set; }
    public TimeSpan PrivilegedProcessorTime { get; set; }
    public float ProcessorTime { get; set; }
    public float WorkingSet { get; set; }
}

i have seen suggestions to compress on client. does it mean i need to convert the object to something like this?

public class PerfData
{
    public long Id { get; set; }    
    public DateTime TimeStamp { get; set; }
    public byte[] CompressedJson{get;set;}
}

appreciate any suggestion and correction. thanks!

Answer 1

The base most class in ServiceStack’s Redis Client is the RedisNativeClient where all operations work on byte[]. The RedisClient is a subclass of RedisNativeClient so you can cast to get the lower-level API.

You would need to compress your value which will end up as a byte[] that you can persist directly into redis. You would need to do the reverse to get it back out again, e.g. retrieve the raw byte[] values and uncompress it.

ServiceStack’s Redis client already has dependency on ServiceStack.Common which contains convenient Stream Extensions to Compress/UnCompress data.