Alerts Table query with time filter update problems

[jcger]

The Alerts table is being fed with a query prop that sets a range of fixed timestamps to get new data (see gte and lte)

As we do not update that filter until reloading the page, the user might experience strange behaviour while using the Alerts Table.
For example, following this workflow:

• User opens a page with our alerts table (for example alerts overview in o11y)
• There is an active alert that will not change status to recovered but keep active next time the rule is executed
• Users stays in that page for longer that it takes to the rule to be executed
• Rule is executed and updates the active alert's timestamp
• Users uses pagination, clicks on reset or does anything that fetches the alerts again
• The active alert will disappear from the list as its timestamp isn't anymore in between the time range that was set in the alerts table when initialising
• If the user reloads, the same alert that disappeared will appear again

Solutions:

Still investigating but our first approach would be to try to update the query that is passed by to the alerts table so it gives us relative timestamp (from now - 15 minutes to now instead of "gte":"2022-09-29T10:44:57.489Z","lte":"2022-09-29T10:59:57.489Z")

Fixing this should close #142053

[elasticmachine]

Pinging @elastic/response-ops (Team:ResponseOps)

[elasticmachine]

Pinging @elastic/actionable-observability (Team: Actionable Observability)