When we need to store information describing the operation of assets at certain moments over a period of time (for example, information from asset sensors), the best choice is a time series storage. Currently, there are dozens of such databases to choose from. The table here features a comparison of the most widely used ones:
|
DalmatinerDB |
InfluxDB |
Prometheus |
Elasticsearch |
Graphite (Whisper) |
|
|
Data types |
float62, int56 |
int64, float64, bool, and string |
float64 |
string, int32, int64, float32, float64, bool, null |
float64 |
|
Data compression ratio |
1 |
2.2 |
1.3 |
22 |
12 |
|
Precision |
millisecond |
nanosecond |
millisecond |
millisecond |
second |
|
Ingest performance (metrics/sec) |
3m |
470k |
800k |
30k |
300k |
|
Data source |
TCP (binary protocol), OpenTSDB (text), Graphite (text), Prometheus (text), Metrics 2.0 (text), InfluxDB... |
