Fixed
Details
Assignee
Yoann RodièreYoann RodièreReporter
Sanne GrinoveroSanne GrinoveroComponents
Fix versions
Priority
Minor
Details
Details
Assignee
Yoann Rodière
Yoann RodièreReporter
Sanne Grinovero
Sanne GrinoveroComponents
Fix versions
Priority
Created November 3, 2016 at 3:27 PM
Updated February 1, 2017 at 11:14 AM
Resolved December 7, 2016 at 1:55 PM
To really get the Jest client usable for production, we need to expose the following configurations:
the pool sizes
the different timeouts on the HTTP requests - it might even be a good idea to set the read timeout a bit higher by default (AFAICS the default for Jest is 3 seconds)
the autodiscovery feature - which allows failover and discovery of new nodes in the cluster: there is a boolean and a frequency: we should at least expose the boolean
To be clear, the autodiscovery only brings "overall" failover, not "per-request" failover. It only implies that failing nodes will eventually be removed from the server list, but a failing request caused by a failing node will not be tried again on another node. Thus will still be valid even if we add autodiscovery.