[MESOS-833] The Status Update Manager should use a back-off mechanism for retried updates.

• Type: Improvement
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 0.16.0
• Labels: reliability
• Assignee: Vinod Kone
• Reporter: Benjamin Mahler
• Created: 21/Nov/13 18:58
• Updated: 26/Nov/13 04:37
• Resolved: 26/Nov/13 04:37

Description

In the event of a delay receiving status update acknowledgements, all status update managers will begin re-sending status updates on a fixed 10 second interval. This can cause a large influx of duplicate status updates in the system, possibly overloading schedulers.

The Status Update Manager should avoid compounding any status update backlog issues through a retry back-off mechanism. Possibly a truncated exponential back-off.