Difference between revisions of "VT Comp Reports"

From EGIWiki
Jump to: navigation, search
Line 1: Line 1:
 
= A new approach to Computing A/R Reports  =
 
= A new approach to Computing A/R Reports  =
  
<br> The EGI Availability and Reliability (A/R) reports are being computed on a monthly basis via a centralized infrastructure that aggregates results from the distributed monitoring. The current implementation of the A/R service cannot be extended to meet the needs of the NGIs and EGI Operations to introduce the new middleware services and regional calculations.
+
The EGI Availability and Reliability (A/R) reports are being computed on a monthly basis via a centralized infrastructure that aggregates results from the distributed monitoring. The current implementation of the A/R service cannot be extended to meet the needs of the NGIs and EGI Operations to introduce the new middleware services and regional calculations.  
  
 
This task will implement a new A/R reporting service using open source project components that will be more flexible and extensible and allow the inclusion of more middleware services into the calculation of A/R metrics and will also add VO-wise metric results (in addition to service-wise, site-wise and NGI-wise provisioning of results). Moreover the profiles under which the calculations are done will be modular and a way to add or remove profiles will be made available and documented. The project is led by GRNET and includes IN2P3 (CNRS) and SRCE.
 
This task will implement a new A/R reporting service using open source project components that will be more flexible and extensible and allow the inclusion of more middleware services into the calculation of A/R metrics and will also add VO-wise metric results (in addition to service-wise, site-wise and NGI-wise provisioning of results). Moreover the profiles under which the calculations are done will be modular and a way to add or remove profiles will be made available and documented. The project is led by GRNET and includes IN2P3 (CNRS) and SRCE.
 +
 +
== Implementation ==
 +
 +
=== Records Collection from PROD MSG Broker network ===
 +
 +
Aggregation of Nagios endpoint results from the EGI Production Message Broker network onto a storage facility upon which the A/R computations will be applied. Filtering of the primary results in a way that only the useful segments from the original information is kept should be applied at this stage so that the computation needs are kept as lightweight as possible.
 +
 +
=== Engine for A/R VO-wise, Service-wise, Site-wise and NGI-wise Computations ===
 +
 +
Development of the SQL queries to be applied on the data the result of which should be the A/R metrics.
 +
 +
=== A/R computation profiles repository ===
 +
 +
Building a repository of computation profiles upon which profiles may be edited, added or removed.
 +
 +
=== A/R API ===
 +
 +
The API will include three basic functionalities: delivery of A/R results, trigger the re-calculation of A/R results (i.e. excluding offending time regions or offending service endpoints from initial calculation) and adding or removal of profiles
 +
 +
=== Interoperation of Operations Portal A/R Frontend module with A/R API ===
 +
 +
Connection of the A/R API with the Operations Portal A/R Frontend module.
 +
 +
== Pilot phase ==
 +
 +
This will validate the new A/R portal results (in comparison to the existing A/R service), it will assess the progress made to date in the new A/R service and document all of the service components
 +
 +
== Service Deployment ==
 +
 +
The service will be packaged, released, documented, rolled out into production and any final adjustment and configuration of the service made as it moves into a production environment. This work and any additional refinements will be undertaken with the EGI OMB

Revision as of 15:19, 9 July 2013

A new approach to Computing A/R Reports

The EGI Availability and Reliability (A/R) reports are being computed on a monthly basis via a centralized infrastructure that aggregates results from the distributed monitoring. The current implementation of the A/R service cannot be extended to meet the needs of the NGIs and EGI Operations to introduce the new middleware services and regional calculations.

This task will implement a new A/R reporting service using open source project components that will be more flexible and extensible and allow the inclusion of more middleware services into the calculation of A/R metrics and will also add VO-wise metric results (in addition to service-wise, site-wise and NGI-wise provisioning of results). Moreover the profiles under which the calculations are done will be modular and a way to add or remove profiles will be made available and documented. The project is led by GRNET and includes IN2P3 (CNRS) and SRCE.

Implementation

Records Collection from PROD MSG Broker network

Aggregation of Nagios endpoint results from the EGI Production Message Broker network onto a storage facility upon which the A/R computations will be applied. Filtering of the primary results in a way that only the useful segments from the original information is kept should be applied at this stage so that the computation needs are kept as lightweight as possible.

Engine for A/R VO-wise, Service-wise, Site-wise and NGI-wise Computations

Development of the SQL queries to be applied on the data the result of which should be the A/R metrics.

A/R computation profiles repository

Building a repository of computation profiles upon which profiles may be edited, added or removed.

A/R API

The API will include three basic functionalities: delivery of A/R results, trigger the re-calculation of A/R results (i.e. excluding offending time regions or offending service endpoints from initial calculation) and adding or removal of profiles

Interoperation of Operations Portal A/R Frontend module with A/R API

Connection of the A/R API with the Operations Portal A/R Frontend module.

Pilot phase

This will validate the new A/R portal results (in comparison to the existing A/R service), it will assess the progress made to date in the new A/R service and document all of the service components

Service Deployment

The service will be packaged, released, documented, rolled out into production and any final adjustment and configuration of the service made as it moves into a production environment. This work and any additional refinements will be undertaken with the EGI OMB