chore: Improve observability

[khorolets]

This PR aims to improve the observability of the node by adding metrics:

• near_lake_block_retry_count_total Total number of retries for storing indexing blocks to S3
• near_lake_block_done_total Total number of indexed blocks

The first can be used to determine whether the service was restarted (e.g., killed by OOM). The second can help determine whether the node is healthy and what the retry rate is.

Additionally, I have added near_lake_build_info similar to the existing near_build_info, but I cannot reuse that since it's private. It will expose the build info. Example:

near_lake_build_info{build="1.37.1",release="0.1.29",rustc_version="1.75.0"}