Confluence Cloud experiencing degraded viewing experience

Summary

On Jun 12, 2026, between 07:04 and 12:10 UTC and on Jun 16, 2026, between 12:56 and 14:35 UTC, a subset of Atlassian customers using Confluence Cloud experienced errors and degraded performance when attempting to view or edit pages. The event was triggered when a database cluster within our infrastructure stopped responding to connection close requests. This caused idle connections to accumulate rather than be released — resulting in increased memory pressure and cascading errors for users. The incident was detected by our automated monitoring within 18 minutes and mitigated by scaling database resources and redeploying, which recycled the idle connections. A permanent fix has since been developed and deployed.

IMPACT

The overall impact was between Jun 12, 2026 07:04 and 12:10 UTC and Jun 16, 2026 12:56 and 14:35 UTC on Confluence Cloud. The June 12 incident caused service disruption to a subset of customers in the EU region while the June 16 incident caused service disruption to a subset of customers in the US region. Impacted users experienced errors or degraded performance when attempting to view, edit, or access Confluence content and spaces.

ROOT CAUSE

1. A database cluster within our infrastructure stopped responding to connection close requests.
2. Idle connections accumulated across clusters rather than being closed, creating a connection leak and increasing memory pressure on the database layer.
3. Hitting max connections on the cluster degraded query performance.
4. Users attempting to view or edit Confluence content experienced errors and slow load times as a result.

REMEDIAL ACTIONS PLAN & NEXT STEPS

We know that outages impact your productivity. While we have a number of testing and preventative processes in place, this specific design limitation in our connection cleanup process was not identified until it was triggered by an adverse infrastructure condition in production.

We have already deployed the following fix:

• Redesigned the database connection cleanup process to prevent unresponsive connections from impacting the closure of others.

To minimize the impact of similar incidents in the future, we will implement additional preventative measures such as:

• Monitoring and alerting for connection cleanup latency and failure rates, so that slowdowns in connection closure are detected early.
• Enhanced safeguards on database connection creation to limit the blast radius of any future connection accumulation events.

We apologize to customers whose services were impacted during this incident. We are taking immediate steps to improve the resilience of our platform.

Thanks,

Atlassian Customer Support