what is the use or purpose of a time-series database?

Redis modules RedisTimeSeries
1

Hi Folks,

It would be great if group members and spare some of their personal experiences on “what is the use or purpose of a time-series database?”

So far, my understanding goes, i would like to add my two penny on this.

Most important thing we need to consider before using time-series database is that our data is structured as:

  1. Series of events.

  2. Samples of one or more values distributed as various snapshots taken in various time.

  3. Metrics collected from different systems/sources over time.

2

Hello Manjeet,

The reason for using a TSDB is the space efficiency and higher speed a dedicated solution can bring over a generic database.

TSDB are meant for streams of data coming from different sources when usually, the writing load is much higher than the reading load. Sensors, for example, send data routinely but querying them usually occurs at lower rate. For specifically, a server might log CPU and RAM levels every 1 millisecond but will update MIN, MAX and AVERAGE levels only every second.

RedisTimeSeries is able to do some aggregations on the server-side which make aggregations queries super fast and simple. It also allows for automated retirement of data so if data was aggregated into the required measurements, the raw data can be erased.If simple API and speed are what you are looking for, you should consider it.

1 Like
3

Hi Ariel,

Thanks for such a good explanation. Specially analogy was awesome. I think i am able to get picture where and how to use Redis TSDB.

Thanks a lot.
Manjeet

4

More info also here: https://redislabs.com/blog/redistimeseries-ga-making-4th-dimension-truly-immersive/

5

Thanks for sharing

6

Good explain!

7

@manjeetkumar53

RedisTimeSeries has features / advantages over others:

  • High volume inserts, low latency reads
  • Query by start time and end-time
  • Aggregated queries (Min, Max, Avg, Sum, Range, Count, First, Last) for any time bucket
  • Configurable maximum retention period
  • Downsampling/Compaction
  • automatically updated aggregated timeseries
  • Secondary index - each time series has labels (field value pairs) which will allows to query by labels