Understanding the rollup tool

[Avi Levi]

I am newbie here and trying to understand blueflood architecture and usage as tsdb. 

Looking at the example below. I see that the same metric (key) have two values on column1 with different values on the "value" column . can you please elaborate what is the purpose of column1 and what is the value stored in "column1" and "value" columns.

Thanks in advance

Avi

1. Example:

                        key  | column1     | value
-----------------------------+-------------+------------------------
tg_tenant.example.metric.one | rollup_type | 0x040842465f4241534943
tg_tenant.example.metric.one |        type |               0x040149

[Shinta Smith]

Hello, Avi,
Our apologies for very late reply.

The example you showed below is the output of the metrics_metadata table. Column1 denotes what kind of metadata it is. The column1=rollup_type metadata is used by Blueflood's rollup process to determine how Blueflood will calculate the rollups. Possible value of this would be BF_BASIC, BF_SET, BF_GAUGE, BF_TIMER, BF_COUNTER (see also RollupType.java class).

The column1=type metadata is no longer used, with the latest master/head version of Blueflood. It used to denote whether or not the metrics are of type String, Boolean or Numeric. We no longer support String and Boolean metrics. Therefore, all metrics are assumed to be numeric. The column1=type should no longer be inserted, if you are using the latest master.

There is also column1=unit metadata. This holds the unit of the metric (for example: seconds, MHz, MegaByte, etc). The value should be whatever string is sent to us by the ingestors.

Hope this helps!

Thanks,
Shinta Smith