NGOP -- What is it?

• Scalable Monitoring System
• "Next Generation OPerations"
• Written at Fermilab
• Written in Python/C++
• Past HEPIX talks discussed architecture etc.
  
  

HEPIX 2002: NGOP UpdateFermilab

What's happened lately?

• Monitoring still more stuff
• Smarter Agents
• Start/Stop standardization
• Performance Monitoring still on hold
• New User Interface code (more to come)
• Impending Lights Out Operation

HEPIX 2002: NGOP UpdateFermilab

More Stuff

• Major Clusters:
  

  Farms	EMail	AFS
  Enstore	WWW	USCMS
  FNALU

• Totals:
    Hosts: ~1,100  Components: ~25,000

HEPIX 2002: NGOP UpdateFermilab

Smarter Agents/Servers

• Ping agent:
    route discovery, outage groups
• Swatch agent:
    combines repeated messages
• Status Engines:
    Delayed effect rules

HEPIX 2002: NGOP UpdateFermilab

New User Interface:

• Split old GUI into:
  • GUI Front End
  • Status Engine(s)
  • Location Server
  
  
• Improves GUI responsiveness.

HEPIX 2002: NGOP UpdateFermilab

Status Engine

• Determines "status" of entities
  -- individual and composite
• Triggers actions on status change
• Multiple GUI's -> One Status Engine
• One Status Engine per "role"
  -- operator, admin, etc.

HEPIX 2002: NGOP UpdateFermilab

GUI Front Ends:

• java gui
  • Responsive, prettier
  • Bigger memory footprint
• web gui
  • FastCGI
  • 

HEPIX 2002: NGOP UpdateFermilab

Demo

• Web GUI
• Web Admin Interface

HEPIX 2002: NGOP UpdateFermilab