Attempt to remedy random CI failures (backport #925)

[mergify]

Summary

#903 lists CI failures that are seen randomly to occur here. I tried running the tests locally a few times and had a couple of failures. After changing the code I did not see any failures. Trying again though on the unaltered code I could not seem to get any failures again.

As such I am not sure the extent, if any, that this will remedy the situation. I added the simuator seed, via options, just in case one of the others, which was not set before, was what ends up getting used so hopefully the extra seeds help. The tests still pass with them set and so they seem to do no harm as such even if they end up not really being used. But lets see how things go. I did not mark this PR to close the random error one i.e. #903 but rather figured, if this seems ok and gets merged, to see how CI goes over time and if we see no re-occurence then close. If I see any random failure as this goes through CI then I think I will close it as clearly it would not be remedying things effectively. (Update: it seems to have gone through CI and passed everything ok, with no random failure, so I guess time will tell how effective this update really can be).

Details and comments

Note: EstimatorOptions does have a seed_estimator and I had initially set that too. But I removed it as it generated a warning when running the tests that it was not used with the local simulator.

---

This is an automatic backport of pull request #925 done by [Mergify](https://mergify.com).