[SLIDER-319] Instrument slider AM with codahale metrics

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: Slider 0.40
• Fix Version/s: Slider 2.0.0
• Component/s: appmaster
• Assignee: Steve Loughran
• Reporter: Steve Loughran
• Created: 14/Aug/14 14:17
• Updated: 16/Mar/15 20:36

Description

Use the codahale metrics lib to instrument slider AM with relevant metrics

1. cluster state, failure state, outstanding requests, ...
2. performance of operations (web &c)

This could then be published:

1. register as JMX provider
2. maybe: publish as ganglia

We should identify and focus on what tends to matter: are the containers up, how many are outstanding, what is the failure rate of the containers

Also: liveness tests?

Issue Links

• is depended upon by: SLIDER-109 Detect and report application liveness
• is related to: SLIDER-768 API to indicate when all component instances are launched
• is related to: SLIDER-772 add sample slider-server.xml with metrics configuration