Cloud outages have a habit of exposing dependencies that are easy to overlook when everything is working as expected.
On 20 October 2026, Google Cloud reported a significant service disruption affecting its London europe-west2 region. The incident was linked to a power event in the europe-west2-b zone, affecting services including Compute Engine, Cloud Storage, Cloud SQL, Kubernetes Engine, BigQuery, Cloud Run and VPC. Customers across the region experienced elevated errors, service availability issues and intermittent network failures.
For businesses, the lesson isn’t that cloud infrastructure is unreliable. It’s that moving workloads to the cloud doesn’t remove the need to plan for disruption.
What happened during the Google Cloud outage?
According to Google Cloud Service Health, a power event caused an outage within the europe-west2-b availability zone. Although the power event was mitigated, Google reported issues across multiple services while workloads recovered.
During the incident, Google advised that, where feasible, affected customers could consider migrating impacted workloads to another available zone or region.
That recommendation highlights an important consideration for any organisation running critical workloads in the cloud: where can your services go when their usual environment becomes unavailable?
Cloud resilience starts with understanding your dependencies
A cloud strategy can look resilient on paper while still relying heavily on a particular region, availability zone, network path or supporting service. That makes understanding dependencies an important part of cloud planning.
Organisations should know where critical workloads are running, what services they depend on and what would happen if part of that infrastructure became temporarily unavailable.
It is worth asking:
- Which applications and workloads are genuinely business critical?
- Where are they hosted?
- Are critical services dependent on a single availability zone or region?
- What happens if that location becomes unavailable?
- How quickly could services be recovered elsewhere?
- When were your recovery and failover processes last tested?
The answers help determine whether your current cloud architecture reflects the level of availability the business actually needs.
Does every business need multi-cloud?
An outage involving a major hyperscaler can quickly prompt conversations about multi-cloud.
That doesn’t necessarily mean every organisation should immediately spread workloads across multiple cloud providers.
Multi-zone, multi-region, hybrid cloud and multi-cloud architectures address different requirements. Increasing infrastructure diversity can improve resilience in some scenarios, but it can also introduce additional management, security and operational considerations.
The better starting point is to understand the organisation’s tolerance for disruption and design accordingly.
A service that can comfortably be unavailable for several hours has very different requirements from an application supporting critical business operations.
Backup and disaster recovery aren't the same as resilience
Having backups is important, but organisations should also understand whether those backups can be recovered effectively during the type of disruption they are designed to protect against.
That means looking beyond the data itself.
Where is it stored? What infrastructure is required to restore it? Which network and identity services does recovery depend upon? Can applications operate from another location if their primary environment is unavailable?
Regular testing matters because a recovery strategy that exists solely in documentation can still leave considerable uncertainty during a real incident.
Look across the entire infrastructure
Cloud resilience also extends beyond the cloud platform itself.
Connectivity, networking, identity, security, data and applications all form part of the environment employees and customers ultimately depend upon.
This is why resilience planning benefits from looking at infrastructure as an interconnected system rather than assessing individual technologies in isolation.
At Cisilion, we help organisations understand those dependencies and design cloud and infrastructure strategies around their availability, security and business continuity requirements.
Is your cloud environment ready for disruption?
The Google Cloud outage is a useful reminder that even large-scale cloud environments can experience disruption. The important question for organisations is how their own architecture responds when that happens.
Understanding where workloads run, identifying critical dependencies and regularly testing recovery plans can give businesses a much clearer picture of their actual resilience.
If recent events have made you reconsider your cloud architecture or business continuity strategy, speak to Cisilion about assessing the resilience of your existing environment.
FAQ
