Improve InfluxDB Logging

[mpass99]

Related to #541

Ensure we receive more details about InfluxDB errors, e.g., in Sentry.

We miss InfluxDB events such as when buffered data points are being dropped.

As seen in #479, the InfluxDB client does not properly provide an interface to such warnings but only logs it to StdOut/StdErr.
The only interface we have is a callback triggered when a HTTP request flushing a batch failed. This callback is not triggered when a batch is dropped because it is expired, because the buffer is full and it got replaced, or because the maximum number of flushing attempts for that batch has been exceeded.

A solution might be to discuss switching to InfluxDB 3.0

• Advantage: We would be up to date and the new Golang client seems to have better error handling [1]
• Disadvantage: The community release of InfluxDB 3.0 is not yet published, but announced for around May. [2]
• Disadvantage: Flux is being deprecated [3].
  • Our Poseidon Dashboard would need to be reimplemented.
  • Following developers would not have to learn the unusual and deprecated Flux language, but could apply their SQL knowledge.