I think we should keep this ticket in mind for T338949: [L] Define SLOs/SLAs for image-suggestions pipelines.

Latest search indices update running now.

ruwiki still has only 68 recommendations

See T345545: Search indices image suggestion tags differ from the dataset used to update.

@mfossati could you update the recovery plan to reflect what is currently happening, as far as I understand the let 2023-08-21's production run compute the proper delta step did not run as expected since you are requesting to import the full dataset in T345545, could you also elaborate on what when wrong that could explain why the delta is not correct?

In T345141#9140348, @dcausse wrote:

as far as I understand the let 2023-08-21's production run compute the proper delta step did not run as expected

That run did generate the expected delta, but see below for what went wrong.

since you are requesting to import the full dataset in T345545, could you also elaborate on what when wrong that could explain why the delta is not correct?

Search’s index update DAG sensor succeeded as soon as that delta was available, leading to a missing delta between the no-ALIS state and that one.
My mistake, I should have elicited the manual update of the in-between delta before unpausing our DAG.
Here’s the sequence of what search indices got:

• no ALIS (broken 2023-08-14)
• good 2023-08-21

As a result, the 2023-08-21 full state would cover the in-between delta:

• the 2023-08-21 delta has already been applied, so all deletions (i.e., __DELETE_GROUPING__) have already happened
• no such rows appear in the full index
• the same subset of 108 k rows that have already been added is also in the full index, but Elastic is expected to skip those identical updates
• the remainder covers the in-between delta, which can be considered as a fix for the broken 2023-08-14 delta