Bucket Stratergy


I am setting InfluxDB 2.0 for my application which runs multiple components:

  • PostgreSQL servers
  • Nginx load balancer servers
  • Application servers (Dockrised)

I plan to run telegraf on each node pushing host metrics (CPU/memory/disk) along with the statistics for each of the specialized services (postgres, nginx, docker) to InfluxDB 2.0
Each server type has multiple nodes.

Would it make sense to log all this data to:

  • A single bucket
  • A bucket per node type (postgres/nginx/docker)
  • 1 bucket per node

Does it make sense/is it possible to segregate buckets by type of data.
Something where host metrics from all 3 types of servers go to one bucket.
Postgres, nginx and docker metrics go to three individual buckets.

Any help here would be very appreciated.

Hello @reshadp,
I recommend putting all of your data related to monitoring your application in one bucket and segregating your data through the use of different measurements and tags. Especially if the data form that nodes have a similar ingest rate.
However, if the ingest rate from those nodes is very different, and you need different retention policies, then I would suggest using multiple buckets.
I recommend taking a look at this documentation:

1 Like

Hi @Anaisdg,
Thanks for your help.
I had read that doc before but need to give it a test with tags.

This topic was automatically closed 60 minutes after the last reply. New replies are no longer allowed.