Increase maturity search buffer to 500ms

[Ariana-B]

Reason for this pull request

When finding a less mature version of a dataset, we previously introduced a 1ms buffer to account for slight timestamp discrepancies. For Landsat 9, this buffer is not sufficient.

Proposed changes

• Increase buffer to +-500ms, giving 1s of wiggle room in finding less mature versions of a dataset

• Closes #xxxx

• Tests added / passed

• Fully documented, including docs/about/whats_new.rst for all changes

---

📚 Documentation preview 📚: https://datacube-core--1458.org.readthedocs.build/en/1458/

[codecov]

Codecov Report

Patch coverage: 100.00% and no project coverage change.

Comparison is base (d8ccecd) 91.82% compared to head (d7e20c9) 91.82%.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #1458   +/-   ##
========================================
  Coverage    91.82%   91.82%           
========================================
  Files          132      132           
  Lines        14516    14516           
========================================
  Hits         13330    13330           
  Misses        1186     1186           

Impacted Files	Coverage Δ
datacube/index/abstract.py	88.28% <100.00%> (ø)

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[SpacemanPaul]

For posterity, the reasoning behind the buffer size is:

1. We have seen time-shifts in reissued Landsat datasets up to a bit over 5ms, so +/- 1ms was clearly not sufficient.
2. The most likely breakdown scenario seems to be satellites in the same constellation (and therefore potentially packaged in the same product) in post-launch tandem orbits for cross-calibration (as occured for example with the Sentinel-3 satellites). The Sentinel-3 tandem orbit had a 30s separation.