Outsystems Outage History

On April 13,2026 between 1:13 and 3:07 AM UTC, an incident occurred that prevented users from logging in to the ODC Portal across multiple regions. The issue was caused by a deployment of a Identity Service that required a database schema change that had not yet been applied, resulting in authentication failures. Service was restored after the team initiated a rollback to the previous stable version. Validation confirmed that login functionality was restored for all users. The incident is fully resolved, and all services are operating as expected.

On March 4th, 2026, between 2:39 am and 10:29 am UTC, ODC experienced an incident that prevented users from accessing the App Detail screen and performing app configurations across all stages. The runtime of existing applications was not affected by this issue. The root cause was a subscription validation misalignment between internal APIs and ODC Portal, which resulted in errors for customers without the ODC Sentry add-on. To prevent future occurrences, we are updating ODC Portal to verify subscription status in alignment with these APIs and improving UI error handling to ensure failures like this do not obstruct the overall page experience.

On January 30th, several ODC customers experienced intermittent issues where the username attribute was missing data from the User entity within applications. The cause was identified as a bug in a recent release of an identity component. This issue prevented specific user and identity provider data from correctly populating the Username Attribute. The incident was resolved by rolling back the affected release. A corrected version has since been deployed to ensure consistent data availability and prevent future occurrences.

On January 27th, we experienced a service disruption in Mentor affecting APAC and EMEA regions. Users encountered elevated error rates and a loss of functionality for Mentor features. The incident was triggered by a third-party service outage, which hosts the models supporting our Data & AI services. While a fallback system was in place, it also became unavailable during the outage. Furthermore, the impact of the external service disruption also prevented our team from implementing a manual configuration fix immediately. We are also working toward an automated fallback mechanism to minimize the need for manual intervention. Additionally, we are implementing agent-level fallbacks to different models at the product level.

On January 22nd, between 15:00 UTC and 17:00 UTC, an incident affected ODC in the Europe \(Frankfurt\) region. Users experienced issues with One-Click Publish, specifically resulting in build timeouts and publish failures. The issue was caused by a communication failure between the internal queuing system and the Build Service, which prevented build processes from completing. Our team resolved the incident by restarting the Build Service and restoring communication. Normal service has been restored, and we are evaluating measures to improve system resilience moving forward.

On January 16th, a service change related to an ongoing migration caused a temporary issue where customers were unable to log in using their custom domains. Our engineering team immediately identified the cause and implemented a mitigation to resolve the issue quickly by temporarily disabling the custom domain redirection feature. A controlled restart of system deployments was performed to expedite this fix. To ensure stability and prevent recurrence, a technical recommendation has been made to enhance the system. This improvement will increase resilience to future updates and configuration changes, helping maintain continuous operation without service interruptions.

On December 8th, we experienced an issue affecting application deployments, resulting in application promotion failures to test and production stages. The incident was caused by an issue related to connection limits on an internal message mechanism. A high volume of retries from a specific Data Services Execution Job, which lacked retry limits, escalated the problem into a connection exhaustion situation. To resolve the issue, a configuration change was implemented to increase the connection limit on the internal message mechanism, and deployments recovered quickly thereafter. To prevent new occurrences, we are creating an action item to limit retries on the offending job to mitigate future connection escalations.

On December 8th, we experienced an issue that affected application publishing. This resulted in users being unable to publish any applications in the Development stage across all regions, encountering the error code OS-DPL-40802 \(Deployment timeout\). The root cause of the incident was a new version of an internal service that was deployed around the same time the problem started. This new version contained changes that caused application deployments to fail. The issue was resolved by reverting the internal service to the previous stable version.

On November 26th, an issue was experienced affecting Personal Editions users, resulting in users seeing a “no permissions” page and being unable to use their Personal Edition environments. Normal service was restored at 21:00 UTC. The incident was caused by the reconfiguration step used to fix an earlier issue. This step correctly re-applied the Identity Provider \(IdP\) but also unintentionally made the affected Personal Editions environments available for re-assignment. As a result, many of the previously fixed Personal Editions were re-assigned to "second" users who were created without any permissions upon first login, leading to the “no permissions” screen. To prevent new occurrences, all Personal Editions environments created during the faulty period that had gone through the reconfiguration were discarded. Impacted users were proactively informed and instructed to request a new Personal Edition.

On November 24, 2025, a new version of our deployment service was released, including an internal library update intended to optimize the processing of external libraries. This update inadvertently caused a logic error that made deployments hang during certain operations, leading to intermittent timeouts for customers deploying external libraries or applications that depend on them. Not all deployments were affected, and the impact varied by region and library volume. We immediately reverted the service to the previous stable version, restoring normal deployment operations. We are working to correct the issue in the internal library and will reintroduce the optimization after thorough validation to prevent future disruptions. We apologize for the inconvenience and appreciate your patience.

On November 18th, we experienced an issue that prevented users with Personal Editions from accessing the ODC Portal login page. Customers who were affected saw the error message: "Something unexpected happened during login. Please try logging in again using a new tab or a different browser". Normal service was restored at 11:09 AM UTC. The incident was caused by a recent system update that introduced an error in how login data was processed. Full access was restored after we reverted the affected system component to its previous version. We apologize for the inconvenience this caused.

A platform-wide issue was identified where log messages generated within successful Timers were inconsistently or completely missing from the monitoring interface. The root cause was a technical race condition during a critical component startup, which prevented logs from reaching the central logging system. Engineering team developed and deployed an update that resolved the issue, and extensive monitoring has confirmed the platform is now stable, with all Timer-originated log entries appearing correctly.

Customers experienced intermittent service issues with core operations, including deployments and authentication, between 2:58 and 10:40 UTC. The disruption was caused by a core system component reaching its maximum connection limit. Our engineering team successfully identified the issue, implemented a configuration update to increase connection capacity, and fully restored service by 10:40 UTC. We are also conducting a full analysis to understand the unexpected usage increase and are enhancing our monitoring for earlier detection of similar events in the future. We apologize for the inconvenience this may have caused.

On November 10th, a system optimization deployment led to a two-hour unavailability of core platform services due to a missing technical dependency. Although quickly resolved, the incident underscored the critical importance of strictly adhering to comprehensive quality assurance gates to ensure platform stability and prevent service disruption.

On October 20, some customers noticed duplicate timer execution logs appearing in the console. A fix was implemented to address this issue. On November 3, another issue occurred in which parts of the timer log display were missing in the console. The investigation determined that the problem originated from a faulty service version introduced in a recent update. The affected version was identified and fixed. The issue was fully resolved on November 7 at 13:07 UTC, and normal log visibility is now available for all customers. We sincerely apologize for the inconvenience and appreciate your patience while we worked to restore full functionality.

On October 30th, between 18:55 UTC and 21:22 UTC, customers experienced degraded success rates and timeouts when using the 1-Click Publish functionality. The issue, caused by an internal system component, was identified and resolved by our engineering team through a service restart, which immediately restored normal operation by 9:22 PM WET.

On October 29th, at approximately 08:15 AM UTC, we observed a significant drop in service success rate primarily affecting Personal Edition instances. Users may have experienced timeout errors. An immediate investigation traced the issue to a recent deployment. Reverting to the previous version resolved the problem, and all affected environments were restored to a healthy condition.

On October 20, some customers across all regions experienced issues logging in using external Identity Providers \(IdPs\), receiving an error or not being redirected to complete authentication. A temporary workaround was provided while the root cause, a faulty service version introduced in recent updates, was identified and reverted. The issue was fully resolved on October 20 at 18:15 UTC, restoring normal login functionality for all customers. The incident duration was extended due to an unrelated AWS outage that impacted service rollback times. We apologize for the inconvenience and appreciate your patience while we restored full functionality.

On October 15th, multiple customers reported being unable to log in to ODC Studio, temporarily interrupting their development work. Access through the ODC Portal remained available. The issue was caused by recent Microsoft Windows updates affecting certain users. A temporary workaround was provided, and uninstalling the updates restored normal access. Microsoft has since released an update to resolve the problem, and customers should now be able to run Windows Update and install it.

On October 7th, 2025, the TimerManager service experienced an issue that prevented it from responding to requests, leading to customers being unable to execute their timers and continue development. The problem was resolved by restarting the affected service, which restored full functionality.