Chunk storage vs Block storage

[sai kiran]

Hi. 

i am pretty new to cortex.. 

We are trying to implement HA and longterm storage for prometheus data. I found couple of options like cortex and thanos.

I found cortex to be interesting and working to set up a POC using cortex. As of now we are planning to use open source backup DB, which points to Cassandra. (may move to cloud soon though)

So my questions are:

1. What is the major difference between block and chunk storages in most of production envs.

2. which is preferred form of storage?

3. How long cortex can hold data and industry standard.

Thank you in advance.

SK

[Marco Pracucci]

Hello!

Please see my reply below.

1. What is the major difference between block and chunk storages in most of production envs.

The chunks storage both require an index store (key-value store like DynamoDB or BigTable) and object store (like S3, GCS, ...), while the blocks storage only require an object store. Also the chunks storage will be soon moved to maintenance mode (we decided it few hours ago during the community call) while blocks storage is under active development.

 

2. which is preferred form of storage?

Blocks storage.

 

3. How long cortex can hold data and industry standard.

As long as you want :)

Marco