v13.0.3 Cluster Flapping Detection

9 views
Skip to first unread message

Evan Tahler

unread,
Feb 18, 2016, 4:44:06 PM2/18/16
to actionHero.js
https://github.com/evantahler/actionhero/releases/tag/v13.0.3

Flapping

When running actionhero startCluster, we we will now detect is the workers/servers under the cluster are rapidly crashing with either an uncaughtException or unhandledRejection. By default, we will check every 30 seconds for this, and if the number of these errors is greater than 2x expected number of workers (set with the --workers flag), the cluster will exit.

Reply all
Reply to author
Forward
0 new messages