Weekend release slots for critical events are important assets that need to be protected. Because there are a limited number of release slots in a year, it is essential that every minute of every event is used efficiently and effectively, maximising delivery capacity. This means that multiple release events have to be scheduled for the same release slot and it needs to be ensured that any conflicting elements are identified, understood and carefully managed.
There is often no system of “air traffic control” to provide a clear, forward-looking view of when projects plan to release. This can lead to conflicting projects inadvertently getting scheduled for release on the same weekend. Release conflicts are usually surfaced through Change Review Boards that are typically convened a week or two before a release slot. However, these reviews are often focused on the stakeholder sign-offs of larger change events rather than the complex inter-dependencies across the release slot. At this late stage, it is incredibly difficult to plan for conflicts without causing problems. This leads to projects being deferred or postponed, increasing costs while delaying the promised value. Therefore, we need to start exploring ways in which we can identify possible conflicts much earlier on in the planning process.
The cost of deferring a release is significant. The knock-on effects to the following release slots create a domino effect impacting the rest of the release calendar. When delayed events are put into later slots, the result is cumulative losses and increasing operational risk and the cost of this inefficiency is rarely tracked or understood. Delays then lead to increased costs; if a project manager gets their whole team geared up to deliver before their release slot and the project is then delayed, they have to go through the whole process again before the next one. This could potentially double the cost of an event while significantly extending the run rate and therefore increasing the cost of the project as a whole.
If conflicts between two simultaneous events are spotted at a late stage, some managers may choose to go ahead with them anyway and hope for the best. Executing projects without fully understanding or planning for possible conflicts between them can exacerbate risks further. This could lead to events failing and negatively affecting customers, increasing regulatory and reputational risk.
A huge amount of planning goes into critical events, yet the forward view of planned change is often managed badly. A greater focus on Enterprise Release Management (ERM) could help to prevent this problem by identifying the content of key release slots in advance. Thinking about releases in terms of content and looking at the joined-up value of multiple releases means clashes can be avoided while making sure essential parts are not left out. Last-minute decisions can lead to important components getting cancelled, making the projects that are delivered redundant because the final piece of the puzzle is missing.
This is a critical gap that needs to be addressed and remains as relevant to the DevOps agenda, in terms of continuous delivery and integration, as it does to the events requiring a higher degree of human orchestration. In most release events these things are not mutually exclusive.