Additional HSDir metrics

Description

At tpo/tpa/team#42057 (comment 3207125), the need for more HSDir metrics was suggested.

Tasks

• Implement the following new metrics for all HS_DESC replies:
  • hsdir_descriptor_not_found_total: tracking the number of times a descriptor query got a NOT_FOUND in a given HSDir.
  • hsdir_descriptor_bad_desc_total.
  • hsdir_descriptor_query_rejected_total.
  • hsdir_descriptor_upload_rejected_total.
  • hsdir_descriptor_unexpected_total.
  • hsdir_descriptor_query_no_hsdir_total.
  • hsdir_descriptor_query_rate_limited_total.
• Create Grafana visualizations for the metrics above (counters, charts and tables).
• Ask around whether would make sense to include regular HSDir checks in implementations such as Arti and Onionbalance, i.e, Onion Service servers to regularly check wheter their descriptors are still alive? Would that make any difference or is this a job for Onionprobe anyway?

Time estimation

• Complexity: small (1 day)
• Uncertainty: low (x1.1)
• Reference (adapted)