Prevent elasticsearch from being disabled on Wikis after some failure
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	Tarrow
	Mar 30 2023, 1:46 PM

Description

Right now it appears that when some platform API jobs fail we have code that then goes and disables elasticsearch for that wiki.
We don't want this to happen; after all we aspire to provide an ES solution for all of our users.
Instead we want to know that something has failed so that we can get on and try to fix it.

There may be more but the following are places where the ElasticSearch backed search state seems to be erroneously changed:
https://github.com/wbstack/api/blame/258eb3ee94bf422b395e07df0d8503acf57e766f/app/Jobs/CirrusSearch/CirrusSearchJob.php#L62
https://github.com/wbstack/api/blame/258eb3ee94bf422b395e07df0d8503acf57e766f/app/Jobs/CirrusSearch/ElasticSearchIndexInit.php#L14
https://github.com/wbstack/api/blame/258eb3ee94bf422b395e07df0d8503acf57e766f/app/Jobs/CirrusSearch/ElasticSearchIndexInit.php#L53

Acceptance Criteria:

No failures should result in changing a wiki's elasticsearch setting state

Related Objects

Mentioned In: T327428: Investigate ways to reduce the resources consumed by ES
T333563: Renable elasticsearch on wikis that have had it disabled
Mentioned Here: T327428: Investigate ways to reduce the resources consumed by ES

Event Timeline

Tarrow created this task.Mar 30 2023, 1:46 PM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptMar 30 2023, 1:46 PM

Tarrow mentioned this in T333563: Renable elasticsearch on wikis that have had it disabled.Mar 30 2023, 2:02 PM

Tarrow updated the task description. (Show Details)Apr 3 2023, 9:48 AM

Tarrow updated the task description. (Show Details)Apr 3 2023, 9:58 AM

Tarrow updated the task description. (Show Details)

Rosalie_WMDE claimed this task.Apr 3 2023, 10:01 AM

Rosalie_WMDE moved this task from Sprint Backlog to Doing on the Wikibase Cloud (WB Cloud Sprint 17) board.

pull request: https://github.com/wbstack/api/pull/595

Evelien_WMDE moved this task from WB Cloud Sprint 17 to WB Cloud Sprint 18 on the Wikibase Cloud board.Apr 12 2023, 11:40 AM

Evelien_WMDE edited projects, added Wikibase Cloud (WB Cloud Sprint 18); removed Wikibase Cloud (WB Cloud Sprint 17).

Evelien_WMDE moved this task from Sprint Backlog to In Review on the Wikibase Cloud (WB Cloud Sprint 18) board.

Deniz_WMDE mentioned this in T327428: Investigate ways to reduce the resources consumed by ES.Apr 12 2023, 12:13 PM

Deniz_WMDE claimed this task.Apr 12 2023, 3:34 PM

I tried testing this PR locally with the Cirrus Patch from https://phabricator.wikimedia.org/T327428#8775552 which previously made the ES initialization of other wikis than the first one fail, and so ES got disabled for them. That doesn't happen anymore with this PR :)

Deniz_WMDE removed Deniz_WMDE as the assignee of this task.Apr 13 2023, 2:14 PM

Deniz_WMDE subscribed.

Deniz_WMDE moved this task from In Review to Waiting for Deploy to Staging on the Wikibase Cloud (WB Cloud Sprint 18) board.Apr 13 2023, 3:01 PM

Tarrow moved this task from Waiting for Deploy to Staging to Blocked/Stalled on the Wikibase Cloud (WB Cloud Sprint 18) board.Apr 14 2023, 10:23 AM

Tarrow moved this task from Blocked/Stalled to Waiting for Deploy to Staging on the Wikibase Cloud (WB Cloud Sprint 18) board.Apr 18 2023, 9:20 AM

Fring moved this task from Waiting for Deploy to Staging to Waiting for Deploy to Production on the Wikibase Cloud (WB Cloud Sprint 18) board.Apr 18 2023, 12:45 PM

Fring moved this task from Waiting for Deploy to Production to Done on the Wikibase Cloud (WB Cloud Sprint 18) board.Apr 18 2023, 1:22 PM

Evelien_WMDE closed this task as Resolved.Apr 26 2023, 12:15 PM

Evelien_WMDE claimed this task.