assess differing counts between 09-29 and 10-30

[paul-butcher]

Having checked diff_tool when deploying the API that points to this pipeline, there are some fairly significant differences in numbers.

e.g. there are more than 300 fewer Images in the new pipeline, but there are also over 1000 more Works.

[paul-butcher]

Blocks #2460

[paul-butcher]

There were a handful of messages on DLQs on both sides. I've re driven them all, but I don't think there were enough to make this kind of difference.

[paul-butcher]

The missing Images are likely to be correct - this is the digmiro suppression in action.

[paul-butcher]

Looking at a few that are present in 09-29 but absent in 10-30, that assumption appears to be correct. a2d2yw3y, a6bttcf3, a6uvdpp2 all correspond to Images in Works that incorrectly showed "View n images" in the selected images section (where n is the wrong number)

[paul-butcher]

I have looked at a few of the Works that are missing in production but present in staging - a24tj3ys, a2vqkjqq, a2x2rumr, a3anrwem, a3trpuce,

They all appear to have fallen down the cracks in the 09-29 pipeline, as they exist in identified but not indexed.

I therefore conclude that 10-30 is better.