ECMWF Newsletter #179

Supporting the WMO Incident Management System

Daniel Varela Santoalla

 

In 2020, ECMWF received a request from the World Meteorological Organization (WMO) to support the establishment of an Incident Management System (IMS) for Regional WMO Integrated Global Observing System (WIGOS) Centres (RWCs). This system was required to manage the observational incidents the RWCs could find within the network of stations under their responsibility. Identification of observational issues, which can concern station metadata (for example incorrect coordinates/altitudes), or data communication/coding, or other reasons, is made by RWCs primarily via the WIGOS Data Quality Monitoring System (WDQMS) webtool (https://wdqms.wmo.int). This is hosted and operated by ECMWF.

ECMWF’s role

ECMWF has been running the Atlassian Jira software as an issue tracking and incident management system since 2011. During this period, we have accumulated expertise in configuration strategies and operational management of the tools provided. Having been granted a so-called ‘open-source licence’ by Atlassian meant that we could afford to use the tools widely without incurring any licensing costs until February 2024.

After the WMO requested it, ECMWF started supporting the WMO in establishing a prototype for a RWC IMS, leveraging our experience in the field with the requirements drawn up by the WMO. Finally, in 2023, the WMO decided to establish their own Jira system. With technical support from ECMWF staff, the IMS for RWCs was migrated to a WMO cloud-based solution, which became operational on 1 February 2024.

Three phases

The collaboration consisted of three distinct phases. During the first, we worked together to design a system that would meet the project's specific requirements. This involved understanding the needs and the roles of RWC and National Focal Points and mapping all those into Jira's technical features, particularly user roles, screens, workflows, priority levels, and custom fields. Roles and the permissions based on them needed to be defined precisely so that only certain users could perform some actions, like opening and closing incidents or commenting on them.

The second phase consisted of the implementation of these requirements and features in an actual Jira project hosted in ECMWF’s infrastructure. This implementation required several iterations with a few internal WMO test users until the design was validated and deemed appropriate for broader testing. This more comprehensive testing required training sessions with members of some specific National Focal Points and Regional WIGOS Centers of some WMO regions to validate the solution and understand any further user requirements that could be of any value for the continuous improvement of the system. One of these sessions, delivered as a webinar, included the National Focal Points of WMO Region III (South America) and the Regional WIGOS Centres for this region, which are Argentina and Brazil. ECMWF participated with the WMO in the delivery of this webinar.

One of the key aspects demonstrated in the webinar was the meaning of the workflow statuses and the conditions under which incidents could be transitioned from one status to another. The figure shows a chart displaying statuses and possible transitions.

WMO RWC Incident Management Workflow.
WMO RWC Incident Management Workflow. The chart shows what happens to observational incidents raised by RWCs.

Once the solution was validated, the third phase of this project involved pre‑operational running of the solution on the existing infrastructure at ECMWF. This pre-operational phase was necessary to validate the solution further and to give WMO staff some time to acquire the necessary expertise to build an infrastructure that could be used to deploy the solution as designed and tested in the previous phases.

Migration to the WMO

Once this infrastructure was ready to take over operations, all data collected during the pre‑operational phase was migrated from the ECMWF infrastructure to the newly built WMO infrastructure. This migration required a good degree of coordination as the data transfer methodology employed required that both origin and destination systems were running the same version of the Jira software. All the hard work came to fruition on 1 February 2024, after which all Regional WIGOS Centres and National Focal Points were instructed to move to the newly deployed system at https://jira.wmo.int.