Difference between revisions of "Procedure Infrastructure Incident Management"

From D4Science Wiki
Jump to: navigation, search
 
(17 intermediate revisions by 4 users not shown)
Line 1: Line 1:
 
__TOC__
 
__TOC__
  
The management of incidents affecting the production infrastructure is based on TRAC issue tracking system. Open incidents can be browsed using TRAC [https://support.d4science-ii.research-infrastructures.eu/report/1 report1 ]. Closed incidents can be browsed using TRAC [https://support.d4science-ii.research-infrastructures.eu/report/2 report 2].
+
== Incident Management==
 +
The management of incidents affecting the production infrastructure is based on RedMine issue tracking system. Open incidents can be browsed using RedMine [https://support.d4science.org/projects/d4science/issues?query_id=207 Open Incident ]. Closed incidents can be browsed using RedMine [https://support.d4science.org/projects/d4science/issues?query_id=208 Closed Incident].
  
 
The resolution of infrastructure incidents is based on the following steps:
 
The resolution of infrastructure incidents is based on the following steps:
Line 13: Line 14:
  
  
'''1) Incident Submission'''
+
==='''Incident Submission'''===
  
Incident tickets are created in [https://support.d4science-ii.research-infrastructures.eu TRAC]. Tickets must be created as follows:
+
Incident tickets are created in [https://support.d4science.org Tracking System]. Tickets must be created as follows:
 
* enter your email address
 
* enter your email address
 
* fill the ''Summary'' text box
 
* fill the ''Summary'' text box
Line 25: Line 26:
  
  
Note: access to TRAC can be requested by emailing to the project [mailto:wp5ATimarine.research-infrastructures-DOT-eu support] mailing list.
+
==='''Prioritization and Assignment'''===
 
+
 
+
'''2) Prioritization and Assignment'''
+
  
 
The [[WP5 Role Infrastructure Manager|Infrastructure Managers]] assign the correct priority to the incident:
 
The [[WP5 Role Infrastructure Manager|Infrastructure Managers]] assign the correct priority to the incident:
Line 34: Line 32:
 
* High: tickets must be solved maximum in 2 working days or in a Maintenance Release
 
* High: tickets must be solved maximum in 2 working days or in a Maintenance Release
 
* Normal: tickets must be solved in one week time or in the next Minor/Major Release  
 
* Normal: tickets must be solved in one week time or in the next Minor/Major Release  
* Low: tickets must be solved in two weeks time or possibly in the the next Minor/Major Release
+
* Low: tickets must be solved in two weeks time or possibly in the next Minor/Major Release
  
Tickets must also be assigned to one member of the [[WP5 Role Support Team|Support Team]]. This is done either by [[WP5 Role Infrastructure Manager|Infrastructure Managers]] or by members of the [[WP5 Role Support Team|Support Team]]. The assigned [[WP5 Role Support Team|Support Team]] member is automatically notified by email. All email notifications are also send to the [mailto:support_team@d4science-ii.research-infrastructures.eu support team] mailing list.
+
Tickets must also be assigned to one member of the [[Role Support Team|Support Team]]. This is done either by [[Role Infrastructure Manager|Infrastructure Managers]] or by members of the [[WP5 Role Support Team|Support Team]]. The assigned [[Role Support Team|Support Team]] member is automatically notified by email. All email notifications are also sent to the [mailto:support_team@d4science.org support team] mailing list.
  
 +
==='''1st Line Support'''===
  
'''3) 1st Line Support'''
+
The assigned member of the [[Role Support Team|Support Team]] analyses the ticket and resolves the issue. Reassignment of the ticket to other members of the [[Role Support Team|Support Team]] might occur as a result of ticket analysis and resolution.
  
The assigned member of the [[WP5 Role Support Team|Support Team]] analyses the ticket and resolves the issue. Reassignment of the ticket to other members of the [[WP5 Role Support Team|Support Team]] might occur as a result of ticket analysis and resolution.
+
If the [[Role Support Team|Support Team]] is not able to close the incident, the ticket is escalated to the Technical Committee (TCom) for further analysis (functional escalation).
  
If the [[WP5 Role Support Team|Support Team]] is not able to close the incident, the ticket is escalated to the TCom for further analysis (functional escalation).
+
==='''2nd Line Support'''===
  
 +
The TCom tries to understand and solve the incident and it may change its prioritization.
  
'''4) 2nd Line Support'''
+
If the TCom is not able to close the incident, the ticket is escalated for a final decision to the management board (hierarchical escalation).
  
The TCom tries to understand and solve the incident.
 
  
If the TCom is not able to close the incident, the ticket is escalated for a final decision to the PEB (hierarchical escalation).
+
==='''Management Decision'''===
  
 +
The management board takes a final decision to allow the resolution of the ticket.
  
'''5) Management Decision'''
 
  
The PEB takes a final decision to allow the resolution of the ticket.
+
==='''Resolution and Recovery'''===
  
 +
The [[Role Support Team|Support Team]] or the TCom solves the incident and recovers the affected service to its normal status.
  
'''6) Resolution and Recovery'''
+
Different steps apply to the resolution of incident tickets, depending on the nature of the incident :
  
The [[WP5 Role Support Team|Support Team]] or the TCom solves the incident and recovers the affected service to its normal status.
+
* To indicate that a problem which required the intervention on the infrastructure has been fixed, the [[Role Support Team|Support Team]] or TCom member should assign to the Progress field the 100% value.
 +
* To indicate that a software problem has been fixed and is now waiting for integration and testing, the [[Role Support Team|Support Team]] or TCom member should assign the milestone "Ready to Release" to the incident ticket and assign to the Progress field the 100% value.
  
To indicate that a software problem has been fixed and is now waiting for integration and testing, the [[WP5 Role Support Team|Support Team]] or TCom member should assign the milestone "Ready to Release" to the incident ticket.
 
  
 +
==='''Incident Closure'''===
  
'''7) Incident Closure'''
+
The [[Role Support Team|Support Team]] closes the ticket according to the ticket priority:
 
+
The [[WP5 Role Support Team|Support Team]] closes the ticket according to the ticket priority:
+
 
* High Priority: tickets can only be closed when the issue has been solved and fixed in the production ecosystem;
 
* High Priority: tickets can only be closed when the issue has been solved and fixed in the production ecosystem;
* Normal/Low Priority: tickets can be closed if the issue was either (1) solved and fixed in the production ecosystem, or (2) fully understood and a new TRAC task created.
+
* Normal/Low Priority: tickets can be closed if the issue was either (1) solved and fixed in the production ecosystem, or (2) fully understood and a new issue ticket created with the tracker 'task' type.
  
'''PLEASE NOTE''' that In case of an intervention on the production infrastructure needed to fix an incident ( e.g. service/GHN restart) , the entitled  [[WP5 Role Support Team|Support Team]]  member should fill the '''Intervention Time''' field of the incident ticket with the info on minutes spent in the infrastructure activity.
+
'''PLEASE NOTE''' that in case of an intervention on the production infrastructure needed to fix an incident ( e.g. service/GHN restart), the entitled  [[Role Support Team|Support Team]]  member should fill the '''Intervention Time''' field of the incident ticket with the info on minutes spent in the infrastructure activity.
  
 
The submitter may re-open the ticket in case he is not satisfied with the ticket response.   
 
The submitter may re-open the ticket in case he is not satisfied with the ticket response.   
  
The [[WP5 Role Support Team|Support Team]] maintains a knowledge database with information about the incident descriptions and resolutions. All closed tickets should be completed with a incident resolution report including the following points: Problem Description, Affected Services, and Resolution Details. This information must be provided as a final comment to the incident TRAC ticket.
+
The [[Role Support Team|Support Team]] maintains a knowledge database with information about the incident descriptions and resolutions. All closed tickets should be completed with an incident resolution report including the following points: Problem Description, Affected Services, and Resolution Details. This information must be provided as a final comment to the incident ticket.
 
+
 
+
== gCube, Hadoop and Runtime  Resources ==
+
 
+
Tickets related to gCube, Hadoop and Runtime resources are followed according to the procedure defined above.
+
 
+
 
+
== gLite Resources ==
+
Tickets related to the usage of gLite resources  are managed according to the procedure defined above. However, the [[WP5 Role Support Team|support Team]] may forward the problem to EGI by creating a new ticket in the EGI [https://gus.fzk.de/pages/home.php GGUS] support service. The [[WP5 Role Support Team|Support Team]] is then responsible for monitoring the tickets submitted in GGUS and update the corresponding TRAC tickets.
+
 
+
Tickets related to the installation and upgrade of gLite are submitted by email to the corresponding EGI [http://www.egi.eu/production-infrastructure/Resource-providers/ NGIs]. These tickets are them processed by [http://www.egi.eu/production-infrastructure/Resource-providers/ NGIs] which either provide an immediate resolution to the problem or forward the problem to the EGI [https://gus.fzk.de/pages/home.php GGUS] support service.
+
 
+
Best effort support and open discussion on any gLite issue is possible using the [mailto:glite-discuss@cern.ch gLite-discuss] mailing list.
+

Latest revision as of 16:31, 5 April 2018

Incident Management

The management of incidents affecting the production infrastructure is based on RedMine issue tracking system. Open incidents can be browsed using RedMine Open Incident . Closed incidents can be browsed using RedMine Closed Incident.

The resolution of infrastructure incidents is based on the following steps:

  1. incident submission
  2. prioritization and assignment
  3. 1st level support
  4. 2nd level support (optional)
  5. management decision (optional)
  6. resolution and recovery
  7. incident closure


Incident Submission

Incident tickets are created in Tracking System. Tickets must be created as follows:

  • enter your email address
  • fill the Summary text box
  • fill the Description text box
  • select the appropriate Application from the list available, if none of the options available apply to your incident please select the "Other" option
  • select the required Priority (optional)
  • add other concerned people in CC (optional)
  • select the appropriate VREVirtual Research Environment. (optional)


Prioritization and Assignment

The Infrastructure Managers assign the correct priority to the incident:

  • High: tickets must be solved maximum in 2 working days or in a Maintenance Release
  • Normal: tickets must be solved in one week time or in the next Minor/Major Release
  • Low: tickets must be solved in two weeks time or possibly in the next Minor/Major Release

Tickets must also be assigned to one member of the Support Team. This is done either by Infrastructure Managers or by members of the Support Team. The assigned Support Team member is automatically notified by email. All email notifications are also sent to the support team mailing list.

1st Line Support

The assigned member of the Support Team analyses the ticket and resolves the issue. Reassignment of the ticket to other members of the Support Team might occur as a result of ticket analysis and resolution.

If the Support Team is not able to close the incident, the ticket is escalated to the Technical Committee (TCom) for further analysis (functional escalation).

2nd Line Support

The TCom tries to understand and solve the incident and it may change its prioritization.

If the TCom is not able to close the incident, the ticket is escalated for a final decision to the management board (hierarchical escalation).


Management Decision

The management board takes a final decision to allow the resolution of the ticket.


Resolution and Recovery

The Support Team or the TCom solves the incident and recovers the affected service to its normal status.

Different steps apply to the resolution of incident tickets, depending on the nature of the incident :

  • To indicate that a problem which required the intervention on the infrastructure has been fixed, the Support Team or TCom member should assign to the Progress field the 100% value.
  • To indicate that a software problem has been fixed and is now waiting for integration and testing, the Support Team or TCom member should assign the milestone "Ready to Release" to the incident ticket and assign to the Progress field the 100% value.


Incident Closure

The Support Team closes the ticket according to the ticket priority:

  • High Priority: tickets can only be closed when the issue has been solved and fixed in the production ecosystem;
  • Normal/Low Priority: tickets can be closed if the issue was either (1) solved and fixed in the production ecosystem, or (2) fully understood and a new issue ticket created with the tracker 'task' type.

PLEASE NOTE that in case of an intervention on the production infrastructure needed to fix an incident ( e.g. service/GHNgCube Hosting Node. restart), the entitled Support Team member should fill the Intervention Time field of the incident ticket with the info on minutes spent in the infrastructure activity.

The submitter may re-open the ticket in case he is not satisfied with the ticket response.

The Support Team maintains a knowledge database with information about the incident descriptions and resolutions. All closed tickets should be completed with an incident resolution report including the following points: Problem Description, Affected Services, and Resolution Details. This information must be provided as a final comment to the incident ticket.