Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @ egi.eu.

Adding Custom Service to Availability Monitoring

From EGIWiki
Revision as of 11:51, 20 February 2013 by Ekarolis (talk | contribs) (→‎Process)
Jump to navigation Jump to search


Introduction

Operational tools such as the GOCDB management system and the SAM monitoring system are key software components for the reliable and stable operation/monitoring of the infrastructure.

GOCDB - Grid Configuration Database (GOCDB) contains general information about the sites participating to the production Grid. Accessed by all the project actors (end-users, sitemanagers, NGI mangers, support teams, VO managers), by other tools and by third party middleware in order to get Grid topology. The portal has a single central installation but a regional package will be developed and deployed on the interested NGIs.

Services registered in GOCDB are described with the following information:

  • Service Type: a unique name that identifies the type of software component deployed on a Grid.
  • Service Endpoint: is a deployed instance of a named service type
  • Endpoint Location: a Service Endpoint may optionally define an Endpoint Location which locates the service (URL).


SAM - The Service Availability Monitoring (SAM) system is used to monitor the resources within the production infrastructure. SAM monitoring data is used for calculation of availability and reliability of grid sites.

The Service Availability Monitoring (SAM) [SAM] system is used to monitor the resources within the production infrastructure. SAM monitoring data is used for calculation of availability and reliability of grid sites. It includes the following components:

  • test execution framework based on the open source monitoring framework Nagios and the Nagios Configuration Generator (NCG)
  • databases which contain topology (gathered from GOCDB and other sources), profiles (mapping between service types and tests), test results and availability and reliability of sites and services
  • visualization portal MyEGI which enables users to access current status, history and availability of monitored sites and services
  • programmatic interface which enables other tools (e.g. Operations Portal, VO dashboards) to access test results and availability and reliability of sites and services
  • probes used to test monitored services which are provided by middleware developers and third parties (e.g. NGIs, Nagios community).


MS421 document (Reference)

Process

1. Register your custom service (hostname / service endpoint) at GOCDB.

Note: if your service does not much any of existing service types at GOCDB, please request new service type . (Service Types descriptions are meaningless to the people who did not had any relations with a middleware e.g. i want to monitor web service which is http, which service type should i choose ? Are there any generic service types available at all ? e.g. http-check). Where are Service Types and Nagios probes mapping ?

2. Choose the set of currently available probes for your service from the list, if you can not find any suitable, you will have to develop the CUSTOM probes for your service. Naming example: CUSTOM.<domain>.<subdomain>.<test_name>, for more examples take a look at Custom Service Types section here. How to develop new probes, please have a look here. After your probes are developed, please follow this procedure to enable them at MyEGI portal.(Probes descriptions are not available, how new people can choose the suitable probes ?)

Note: if your service has one hostname / service endpoint and you would like to test multiple functionality e.g. http-check and db-check, please make sure that another custom service you will add in the future for the same created service type also provides same multiple functionality, because service types are associated with a set of probes.


Service provider Operations
1. Register service endpoint at GOCDB.
1. Validates the service endpoint registration.
2. Find the suitable service type at GOCDB.

2.1 If your service does not much any of existing service types at GOCDB, please request new service type .

2. If existing service type was chosen, the rest of operations are automatic.

2.1 If new service type was requested, operations will assess it through OMB and OTAG.

3. If existing suitable service type was chosen at GOCDB, please set option "Monitoring" to start receiving probes.

3.1 If new service type was requested choose the set of currently available probes for your service from the list.
3.2 If you can not find any suitable probes, you will have to develop the CUSTOM probes for your service. Naming example: CUSTOM.<domain>.<subdomain>.<test_name>, for more examples take a look at Custom Service Types section here. How to develop new probes, please have a look here. After your probes are developed, please follow this procedure to enable them at MyEGI portal.

3. Rest of operations are automatic.

3.1 If new service type was requested and existing probes in SAM framework were chosen, they have to be enabled in a profile of local NGI based on service location
3.2 If new probes were submitted, operations will start this procedure.