fix: replication allocation invocation timeout

[alanshaw]

This PR increases the validity time of the blob/replica/allocate invocation so that storage nodes have a bit more time to use it as proof when requesting a blob/retrieve delegation from the node storing the data.

I am seeing replica nodes fail to transfer data because of this:

2025-10-29T15:50:15.856Z        WARN    replicator      worker/worker.go:309    Error running job, retrying     {"queue": "replication", "name": "transfer-task", "attempt": 1, "max_attempts": 10, "error": "failed to accept replication source blob 12203b
e57a446cca42697381285855e1ec0f3fe15322b7e56aab12315122e1ed4633: requesting blob/retrieve delegation: Claim {can:\"blob/replica/allocate\"} is not authorized\n  - Proof bafyreibgcaoeeu5pzra73d5lc3ryx55uyae6bduyal5ggxn7wet6oypkti has expired on 2025-10-29
T15:50:00Z"}