Use monotonic clock for timeouts; harden db-sync progress wait

[mkoura]

Use monotonic clock for durations and deadlines

Replace time.time() with time.monotonic() where values are used only for
elapsed-time measurement or in-process deadlines, so timeouts are immune to
wall-clock jumps (NTP sync, manual adjustment):

• dbsync_utils: retry_query timeout and db-sync progress poll
• test_tx_many_utxos: UTxO generation duration
• test_blocks: epoch-deadline loops; variables renamed from *_timestamp to
  *_deadline as they no longer hold wall-clock timestamps

Wall-clock time.time() is intentionally kept where timestamps are persisted,
shared across processes or compared to file mtimes (logfiles offset and
ignore-rule files, status db, tx-name generation).

Harden wait_for_db_sync_completion

While converting the clock usage, review uncovered real bugs in the polling
logic:

• query_db_sync_progress crashed with TypeError on an empty block table:
  the aggregate query returns a (None,) row, so float(None) raised and the
  else 0.0 branch was dead. This was hit routinely, as the only caller
  (restart_with_config) recreates the database right before polling. The
  function now returns None in that case (annotated float | None).
• A None progress is now retryable: the initial query is retried by
  retry_query, and the poll loop keeps polling with the last known progress
  until the deadline. A genuine 0.0% progress no longer aborts the wait.
• The loop logs the value it compares instead of issuing a second redundant DB
  query per iteration.
• Timeout raises DbSyncTimeoutError (a TimeoutError subclass) for module
  consistency and reports actual elapsed time.
• Docstring fixes: hardcoded 99% vs the expected_progress parameter, typo.

[Copilot]

Pull request overview

This PR updates timeout/deadline handling to use monotonic time (avoiding wall-clock jumps) and hardens db-sync progress polling so empty databases and None progress are handled correctly without aborting prematurely.

Changes:

• Switched elapsed-time and deadline calculations from time.time() to time.monotonic() in db-sync utilities and relevant tests.
• Updated db-sync progress querying to return None when the block table is empty, and adjusted the polling loop to treat that as retryable/pollable.
• Standardized db-sync timeout behavior to raise DbSyncTimeoutError (a TimeoutError subclass) and improved progress logging behavior.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
cardano_node_tests/utils/dbsync_utils.py	Uses monotonic deadlines for retries/polling and improves db-sync completion waiting logic and error semantics.
cardano_node_tests/utils/dbsync_queries.py	Returns None (instead of crashing) when sync progress cannot be computed due to an empty block table.
cardano_node_tests/tests/test_tx_many_utxos.py	Uses monotonic timing for reporting UTxO generation duration.
cardano_node_tests/tests/test_blocks.py	Uses monotonic deadlines for epoch/test loops while keeping wall-clock time for persisted tx name generation.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.