fix: add targeted logfire span for PDS record fetching (#721)

The album endpoint was showing ~205ms average but HTTP Request spans
to PDS were showing 0ms duration - an instrumentation gap.

This adds a surgical `logfire.span()` wrapper around `get_record_public()`
to capture actual PDS HTTP call duration without flooding spans like
global httpx instrumentation would.

Span attributes:
- collection: the ATProto collection being fetched
- rkey: the record key
- pds_host: the PDS hostname

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.5 <[e-mail]>

backend/src/backend/_internal/atproto/records/fm_plyr/track.py

···
         Exception: if fetch fails
     """
     import httpx
+    import logfire
 
     repo, collection, rkey = parse_at_uri(record_uri)
 
···
     url = f"{base_url}/xrpc/com.atproto.repo.getRecord"
     params = {"repo": repo, "collection": collection, "rkey": rkey}
 
-    async with httpx.AsyncClient() as client:
-        response = await client.get(url, params=params, timeout=10.0)
+    with logfire.span(
+        "pds_get_record {collection}",
+        collection=collection,
+        rkey=rkey,
+        pds_host=base_url.replace("https://", ""),
+    ):
+        async with httpx.AsyncClient() as client:
+            response = await client.get(url, params=params, timeout=10.0)
 
     if response.status_code != 200:
         raise Exception(