On Tuesday evening, the Independent Electoral and Boundaries Commission (IEBC) issued a brief statement announcing a temporary disruption to its online voter services. The commission noted it was undertaking a routine ICT upgrade to enhance reliability and security. Consequently, the public voter search portal, online voter transfers, and requests to change registration details were taken offline, leaving citizens staring at digital dead ends in the middle of a standard workweek.
In 2026, this approach to critical civic infrastructure is not just inconvenient; it highlights a profound architectural deficit. In an era of high-availability cloud computing, pulling entire public-facing databases offline for routine maintenance is a relic of an older tech stack.
The private sector solved this problem over a decade ago.
Consider how Kenya’s financial giants handle critical systems. When Safaricom or major retail banks undergo massive core infrastructure upgrades, the public rarely notices. When downtime is unavoidable, they follow a highly disciplined playbook.
- The Midnight Window: Upgrades are executed when user traffic hits rock bottom, typically between 1:00 AM and 4:00 AM on a Sunday morning. The IEBC’s decision to disrupt civic services during the mid-week peak stands in sharp contrast to basic service-level standards.
- Surgical Precision: Private enterprises do not issue open-ended notices. They give exact timeframes, such as stating that services will be unavailable from 2:15 AM to 3:45 AM. A vague, open-ended temporary disruption notice from an electoral body creates an unnecessary vacuum of information.
For election-related data, any disruption should be aggressively mitigated. If a system must go down, it should be heavily documented, tightly scheduled, and compartmentalized.
From a modern software engineering perspective, taking down a read-heavy public verification portal for an upgrade points to architectural bottlenecks. A resilient public-sector stack should leverage modern cloud-native strategies to ensure uninterrupted service.
1. Decoupling reads from writes (CQRS)
Voter verification is almost entirely read-only traffic. By implementing Command Query Responsibility Segregation (CQRS), the public-facing search portal could be entirely decoupled from the primary transactional database. Citizens would query highly available Read Replicas distributed across multiple zones. Even if the primary database is taken offline for a backend overhaul, the public search portal remains active, serving cached or replicated data without a second of downtime.
2. Blue-green deployments
Announcing a service blackout implies that code or infrastructure changes are being pushed directly onto a single, live production environment. A modern approach utilizes Blue-Green Deployments, where an identical, isolated environment (Green) is spun up next to the live production version (Blue). The engineering team applies the upgrades and runs comprehensive testing on the Green environment. Once verified, traffic is seamlessly routed over via a load balancer. If an unexpected error occurs, traffic rolls back instantly to ensure total downtime remains at zero.
3. Change data capture (CDC) over bulk Loads
The days of freezing databases to run massive, manual batch processing updates or bulk data migrations are over. Modern data pipelines use Change Data Capture (CDC) frameworks, like Apache Kafka or Debezium, to stream transactional updates in real time. Infrastructure adapts incrementally in milliseconds, eliminating the need for administrative freezes.
4. Breaking the monolith
Because voter search, transfers, and changes of particulars went down simultaneously, it is highly likely these services are bound within a single monolithic architecture. Shifting to decoupled microservices ensures that a logic change to the voter transfer service does not inadvertently blind the voter search portal.
Beyond the technical architecture, there is a looming transparency issue. The IEBC’s statement relied on standard corporate boilerplate, claiming they were undertaking a routine upgrade of their ICT systems to enhance reliability, security, and performance.
In an environment where electoral integrity is under constant scrutiny, vague phrasing breeds skepticism and invites conspiracy theories. Transparency is not just good PR; it is a core security feature of democracy.
The IEBC should treat the Kenyan electorate like tech-literate stakeholders. Instead of hiding behind generic tech terms, a transparent commission would explicitly state what is being changed: “We are migrating our database indexing to a distributed cloud architecture to handle peak traffic for the upcoming cycle, and upgrading our data-at-rest encryption protocols on voter profile fields.”