[Misc] Add opentelemetry metrics to CTA

Description

This MR adds opentelemetry to CTA.

• Some small abstractions have been introduced that allow developers to:
  • Create a telemetry configuration
  • Initialise telemetry in a consistent manner using this configuration
  • Globally (uniquely) define instruments per library.instrument registrations are considered a semantic error
• A number of different metrics:
  • DB metrics:
    • Histogram of query duration (includes possible errors)
    • UpDownCounter for number of active DB connections
  • Frontend metrics:
    • Histogram of the processing duration of frontend requests
  • Scheduler metrics:
    • Histogram of the latency of lock requires (i.e. time to wait for the lock to be acquired)
    • Counter for number of requests enqueued
  • Taped metrics:
    • Counter for number of files transferred
    • Counter for number of mounts
    • UpDownCounter for number of tape read/write threads
    • UpDownCounter for number of disk read/write threads

Note that any histogram info can also be used to extract information about counts/rates

Overview of the changes in the MR:

• .gitlab/ci: manually triggered testing for testing telemetry correctness
• common/telemetry/: basic code to provide telemetry functionality
• continuousintegration/: add cta-dependencies as a repo as this contains the opentelemetry-cpp RPMs. Update the pods to use a flexible telemetry configuration and add basic system test for telemetry correctness. Add option --enable-telemetry to build_deploy.sh
• The remainder is for the initialisation of telemetry and the usage of instruments

Checklist

• Documentation reflects the changes made. See https://gitlab.cern.ch/cta/eoscta-docs/-/merge_requests/73
• Merge Request title is clear, concise, and suitable as a changelog entry. See this link
• Create operations ticket for setting this up and testing this in preproduction. Done, see https://gitlab.cern.ch/cta/operations/-/issues/1784

References

Closes #1135