Implement rate limiting

[Teneroy]

Reasoning and Description:
In order to prevent high spendings and hitting AI core rate limiting, we need to introduce our own rate limiting per cluster which uses our companion.

Tasks:

1. Make sure our traces in Langfuse have identifiers of clusters which uses our companion
   Whenever we receive a request we need to:
2. Pull number of tokens used by cluster from 00:00UTC to 23:59UTC (possible through Langfuse API) or within 24h period(whatever is easier).
3. If the number of tokens is higher than a constant we set, return a message to the user that they over consumed their token usage and should come back after 23:59UTC or after X minutes(if we go with 24h approach)

Acceptance criteria:

• Traces are being identified per cluster
• Total token usage is being pulled within agreed time range and compared against the constant

[Teneroy]

https://api.reference.langfuse.com/#get-/api/public/observations