postmortem.io Beta
Incident

Impaired Cluster Operations – AWS me-central-1 (United Arab Emirates) and AWS me-south-1 (Bahrain)

MongoDB Atlas Read the original status page ↗

Impaired Cluster Operations – AWS me-central-1 (United Arab Emirates) and AWS me-south-1 (Bahrain)

MongoDB Cloud Status incident, opened 2026-03-01, status: monitoring

Update (2026-03-01, identified): Impact: Some Atlas customers with deployments in the AWS me-central-1 (United Arab Emirates) region may experience delays in cluster operations (for example, creating, scaling, or modifying clusters and infrastructure-level maintenance tasks).

Root cause: AWS is reporting connectivity and power issues affecting APIs and instances in a single Availability Zone in the me-central-1 Region, which can cause increased errors and latency for some AWS services and workflows. For the latest details from AWS, see the AWS Health Dashboard.

What you might see:

  • Cluster create/scale/modify operations remaining in a pending or in-progress state longer than usual
  • Slower completion of some maintenance and infrastructure operations in the affected region

Our actions:

  • Monitoring the AWS incident and Atlas fleet behavior in me-central-1

Customer action: No action is required at this time. We expect delayed operations to complete as AWS resolves the underlying issue. If you have time-sensitive changes that are blocked and need urgent assistance, please contact MongoDB Support.

Update (2026-03-01): One or more instances backing MongoDB Atlas clusters in the AWS ME-CENTRAL-1 (United Arab Emirates) region may have been temporarily unavailable due to an underlying AWS issue in a single Availability Zone. According to AWS, a localized power issue in this Availability Zone has affected connectivity and power for some APIs and instances, and may have caused increased errors, latency, or unavailability for certain workloads.

Update (2026-03-02): According to AWS, two Availability Zones in me-central-1 (UAE) and one Availability Zone in me-south-1 (Bahrain) continue to experience outages.

CURRENT IMPACT: Atlas clusters deployed only in me-central-1 may be fully unavailable. For clusters operating solely in me-central-1, backup operations in this region are impaired. Clusters and backups should remain operational in me-south-1 at this time. Clusters in me-south-1 may be running at reduced capacity due to zonal constraints. AWS has indicated an extended resolution timeline for service in the impacted Regions and Availability Zones.

RECOMMENDED CUSTOMER ACTIONS: If you have a multi-region configuration that includes a healthy secondary region, we recommend routing application traffic to that secondary region. If you have clusters in me-south-1 taking backups, we recommend enabling “Additional Snapshot Copies Policy” to a different region.

WHAT WE ARE DOING: We recognize the severity of this situation and the impact it may have on your workloads. The situation continues to evolve, and we are closely monitoring AWS status and regional health signals. We are validating Atlas behavior across affected and unaffected regions, and working to mitigate customer impact where possible.

Update (2026-03-02, later): AWS is currently experiencing infrastructure impairments in two Middle East regions:

  • ME-CENTRAL-1 (Middle East – Central): Two Availability Zones are impaired due to an underlying power and facilities issue. This is resulting in elevated error rates and latency across several AWS services. Recovery is currently expected to take at least a day, per AWS.
  • ME-SOUTH-1 (Middle East – Bahrain): A single Availability Zone is experiencing a localized power issue. While some services have shifted traffic to unaffected Availability Zones, resources deployed in the impacted zone may continue to experience availability issues, elevated error rates, or extended recovery timelines. Recovery is also currently expected to take at least a day.

Because MongoDB Atlas runs on AWS infrastructure, Atlas clusters deployed in the affected Availability Zones may be unavailable or experience operational impact, and certain Atlas management operations (including the Atlas UI, API, and CLI) may show increased error rates during this time.

The impact of these events depends on your Atlas cluster topology and configuration:

  • Customers using multi–region deployments, global clusters, or replicas located in unaffected Availability Zones or regions may see little to no impact.
  • Customers with clusters or workloads primarily located in the affected Availability Zones may experience reduced availability or degraded performance.
  • Clusters deployed entirely in unaffected Availability Zones or outside the impacted regions are operating normally.

AVAILABLE OPTIONS: Depending on your application requirements and cluster configuration, you may consider the following options: (1) Move the cluster to another region that is not impacted. (2) Add electable nodes in another region, which can enable the cluster to establish a minimum viable topology and restore database operations. (3) Take no action and wait for the affected Availability Zones to be fully recovered.

Updates through March, April, and May 2026: Two Availability Zones in me-central-1 (UAE) and one Availability Zone in me-south-1 (Bahrain) continued to experience impairments over the following weeks. MongoDB continued to assist customers with remediation, encouraging customers running workloads in the Middle East to consider alternate Atlas regions in North America, Europe, or Asia Pacific, and to extend backup snapshot expirations in the affected regions as needed. The update cadence moved from daily, to every few days, to weekly, and eventually to monthly as the situation stabilized without full recovery.

By late March 2026, MongoDB reported that clusters in me-south-1 could no longer be modified and might be unable to process normal operations, and strongly recommended that customers with workloads in the affected regions move their clusters to alternate regions rather than wait for full availability to be restored.

Update (2026-04-01): Two Availability Zones in AWS me-central-1 (UAE) continue to experience significant impairments. In addition, workloads in me-south-1 (Bahrain) are currently not operational, cannot be modified, and are not able to migrate to other regions or support other disaster recovery actions such as backup and restore. We strongly recommend that customers with workloads in AWS me-central-1 act now to move their clusters to alternate regions instead of waiting for full availability to be restored. There are currently no recovery options available for data that is only present in the AWS me-south-1 region.

Update (2026-05-06): AWS me-central-1 (UAE) and me-south-1 (Bahrain) have suffered damage in these regions. Workloads in me-central-1 (UAE) continue to experience significant impairments. In addition, workloads in me-south-1 (Bahrain) are currently not operational, cannot be modified, and are not able to migrate to other regions or support other disaster recovery actions such as backup and restore. According to the AWS Service Health Dashboard update from April 30th, recovery is expected to take several months.

Update (2026-06-03, most recent): AWS me-central-1 (UAE) and me-south-1 (Bahrain) have suffered damage in these regions. Workloads in me-central-1 (UAE) continue to experience significant impairments. In addition, workloads in me-south-1 (Bahrain) are currently not operational, cannot be modified, and are not able to migrate to other regions or support other disaster recovery actions such as backup and restore. We continue to assist customers in their chosen remediation path to the extent possible. We will move to an as-needed status post update cadence as new information becomes available.