elastic/elasticsearch

ETA for Reindex API

Open

Aperta il 25 ago 2021

Vedi su GitHub
 (2 commenti) (0 reazioni) (0 assegnatari)Java (76.700 star) (25.882 fork)batch import
:Distributed/Reindex>enhancementTeam:Distributedhelp wanted

Descrizione

The get task API (e.g. GET _tasks/lOHn4ATPST6LfKCW6m539w:517820) should include an estimated time of completion where applicable.

Specifically when using the reindex API, I had to write myself a basic tool some time ago to estimate when the task would be completed using the total, created, updated, start_time_in_millis, and running_time_in_nanos properties. (Although I believe in cases where the reindex operation omits some documents, this may not even be sufficient!)

It would be great if this functionality was built into this API.

Guida contributor

Tech stack
javarest api
Dominio
apisearchobservability
Tipo issue
feature
DifficoltàQuanto dovrebbe essere impegnativa per un nuovo contributor.
4
Tempo stimatoTempo stimato per completare e verificare uno scope piccolo.
over 1 week
Stato attivitàQuanto è probabile che la issue sia ancora attiva e reviewable.
stale
ChiarezzaQuanto chiaramente la issue descrive problema, scope e risultato atteso.
mostly clear
Prerequisiti
Elasticsearch task frameworkREST API endpointsJava concurrencyTime estimation logic
Adatta ai principiantiQuanto la issue è adatta a contributor alla prima esperienza.
20
Direzione di ricerca
Investigate the Elasticsearch task management system, specifically the get task API implementation at `GET tasks/{taskId}`. Understand how the reindex operation tracks progress using fields like `total`, `created`, `updated`, `start time in millis`, and `running time in nanos`. Research potential algorithms like exponential moving average to estimate ETA, considering that documents may be omitted. Review existing discussions in the issue and any linked resources. Look at the `TasksClient.java` and `TransportTasksAction.java` source files for implementation details. Proposing adding an `estimated completion time` field to the task response is a starting point.