2022 – 2023 · Montpellier, France · Hybrid
Project officer (infrastructure operations) — employee savings
CGI — Leader in employee savings plan administration
CGI runs several engagements for this client: a fifteen-person application maintenance carrying releases and changes, a ten-person Siebel migration, and the operational upkeep of the machine estate under CGI's responsibility — non-production, pre-production and production environments. The project manager carrying the whole being staffed beyond a hundred per cent, the scope was split into domains, each entrusted to a project officer. This one covers operational upkeep, the smallest of the three.
Two strands. Day to day, handling alerts and provisioning the machines the development teams need, so that infrastructure never becomes the engagement's bottleneck. Every fortnight, a technical committee gathering some thirty participants around an infrastructure and blocking-points backlog. Most of the operational upkeep is sold as an annual fixed price; the workstreams that fall outside it get their own study, estimate and schedule.
Two self-sufficient half-time resources on monitoring and escalation, and a management load of around twenty per cent. Hosted in the client's data centre, virtualisation by hypervisor; development chain on GitLab, Jenkins, SonarQube, Docker and Nexus. The engagement stopped in September 2023: by the summer, assignments totalled more than a hundred per cent between product ownership, a project management role ramping up at an omnichannel retailer, the tail end of a rail maintenance engagement and an industrial eco-design assessment.
Achievements
- 40 virtual machines across eight environments
Operational upkeep of the estate
Handling alerts and incidents across the scope under CGI's responsibility — non-production, pre-production and production environments — with two self-sufficient resources on monitoring and escalation. Provisioning the machines the development teams need, so that development capacity is never constrained by infrastructure.
- 30 client and CGI participants
Fortnightly technical committee on an infrastructure backlog
Running a committee every fortnight on some thirty topics held in a kanban: security workstreams, estate management, dependencies with the application maintenance and with the client's projects. Some thirty participants — infrastructure and security leads, project managers and release managers, on both the client and CGI sides. Presenting the topics, an open discussion on progress, actions recorded, then the kanban extracted into minutes: actions under way, actions to come, state of play since the previous committee.
Infrastructure workstreams priced outside the fixed price
Beyond the annual fixed price, each major workstream gets its own study, quote, approval and schedule. Carried out over the period: moving the virtual machines from Windows 7 to Windows 10, with remastering and data migration on developer workstations; installing the new firewall across the whole estate under CGI's responsibility; verifying and running the disaster recovery plan; analysing and fixing faults in the continuous integration and deployment chain.
Annual re-evaluation of the service envelope
Recording the two resources' daily activity in an internal tool, year after year, in order to re-evaluate the operational upkeep fixed-price envelope.
Skills applied
- Application maintenance and operations management
- DevOps
- Disaster recovery plan (DRP)
- Excel
- Fixed-price estimating
- InfraOps
- Jira
- Project financial tracking
- Project management