Best practices for failsafe influxdb operation

Hi community,
I plan to offer a commercial service (public sector, OSS is mandatory). It includes several wireless (LoRaWAN) sensors sending their data (via telegraf) into an influxdb.

I want to assure as much as possible that:

  • I don’t loose any data in the database
  • all sensor data is written to the database
  • database is always available to read data by downstream tools (e.g. Grafana)

Currently I’m doing daily backups of the database and store it safely.
But maybe I should do more. Some thoughts:

  • Running 2 instances of influxdb on different servers and let telegraf fill both databases with the same data?
  • Instead of sending the same data to both instances of influxdb, copy (e.g. ones a day) all data from one database to the other?
  • Kind of failover strategy between both influxdb instances?

I’m not sure what is actually helpful or worth the effort. I was not able to find documentation about this topic in the influx docs.
Do you have any advice or a link to some information around this topic?

Have you considered ‘Edge Data replication’ ? Highly recommend looking into it : InfluxDB Edge Data Replication | InfluxData and here Replicate data from InfluxDB OSS | InfluxDB OSS v2 Documentation

I haven’t even heard the term :sweat_smile:!
I will look into your linked documents. Thank you very much.

1 Like