Historique des pannes de Aptible

1. 22 octobre 2026 à 00:24 UTCresolved

   Nous sommes au courant d'un incident de sécurité impliquant des certificats eMudhra Certificate Authority (CA) compromis qui pourraient affecter la sécurité HTTPS sur les systèmes avec des magasins de confiance de certificats dépassés. La plateforme Apible n'est pas directement touchée par ce problème. Cependant, par précaution, nous recommandons aux clients: - Mettre à jour les systèmes d'exploitation et les navigateurs Web à leurs dernières versions de sécurité pour s'assurer que les magasins de fiducie certificats sont à jour. - Mise à niveau vers CLI acceptable v1.0.5, qui comprend des mesures d'atténuation supplémentaires pour les systèmes dont les magasins de fiducie OS n'ont pas encore été mis à jour : https://www.aptible.com/docs/reference/aptible-cli/overview Pour plus d'information, voir l'avis de sécurité de Cloudflare : https://blog.cloudflare.com/emudhra-certificat-compromis

Traduit automatiquement depuis la mise à jour officielle de l'incident.

1. 27 juillet 2026 à 10:31 UTCinvestigating

   AWS connaît un problème avec leurs API dans nous-est-1, et de nombreux services sont touchés à ce moment. Cela peut conduire à des opérations ratées. Pour éviter d'autres problèmes, nous avons des opérations pour personnes handicapées dans cette région tout en comprenant l'impact total.

2. 27 juillet 2026 à 12:55 UTCinvestigating

   Les opérations demeurent désactivées en nous-est-1 pendant que nous continuons à enquêter sur les problèmes de réseautage des SSFE dans la région. Nous testons les mesures d'atténuation et rétablissons les opérations dès qu'il est sûr de le faire. Nous fournirons une autre mise à jour au plus tard à 10 h, heure de l'Est.

3. 27 juillet 2026 à 13:43 UTCmonitoring

   Nous avons appliqué une mesure d'atténuation et des opérations réactivées en nous-est-1. Nous veillons à la stabilité et résoudrons cet incident une fois que nous aurons confirmé que les conditions demeurent stables.

4. 27 juillet 2026 à 15:12 UTCresolved

   Cet incident a été résolu.

Traduit automatiquement depuis la mise à jour officielle de l'incident.

1. 16 juillet 2026 à 22:31 UTCidentified

   Nous sommes au courant des demandes reportées et reportées à la passerelle LLM. Nous avons apporté certains changements afin de minimiser l'impact tandis que les ingénieurs continuent de s'attaquer à la cause profonde.

2. 17 juillet 2026 à 00:02 UTCmonitoring

   Nous avons accru la capacité de ressources et nous suivons les résultats.

3. 17 juillet 2026 à 12:14 UTCmonitoring

   Après une régression en performance à partir de 9:57 UTC (5:57 ETP) et jusqu'à 11:12 UTC (7:12 ETP), nous avons déployé une correction potentielle et mis à jour notre échelle de service. Nous continuons de vérifier si ces changements résolvent complètement le problème avant de clore cet incident.

4. 17 juillet 2026 à 17:04 UTCresolved

   Suite aux corrections déployées lors de la mise à jour précédente, nous avons vu des performances stables sur la passerelle LLM depuis 7:12 HAE et sommes en train de résoudre le problème. La cause principale de ce problème était une grave dégradation des performances dans trois paramètres spécifiques: • "GET/modèles" • `GET /v1/modèles` • `GET/metadata/modèles` ... qui renvoient une liste de modèles disponibles. Le problème de rendement sur ces paramètres était suffisamment grave pour causer une file d'attente importante et une dégradation de la performance pour d'autres paramètres non affectés. La correction que nous avons déployée a ajouté de la mise en cache sur ce paramètre pour assurer des réponses rapides. Puisque ces corrections ont été déployées, nous ne voyons plus de file d'attente de requêtes ou d'impacts sur d'autres paramètres de l'API de la passerelle LLM.

Traduit automatiquement depuis la mise à jour officielle de l'incident.

1. 8 juin 2026 à 18:28 UTCinvestigating

   As part of our security policy and ongoing commitment to protecting customer workloads, we are applying patches for a newly disclosed Systemd vulnerability (CVE-2026-40226) across our platform. This patching will be rolled out automatically over the next several days. During this process, you may experience a momentary network disruption to running applications and databases. Customers configured for high-availability with horizontally-scaled services should experience little to no disruption from this update. No action is required on your part. If you have any questions or concerns, please reach out to our Support team at https://contact.aptible.com/.

2. 23 juin 2026 à 21:48 UTCresolved

   This incident has been resolved.

1. 3 juin 2026 à 16:02 UTCinvestigating

   There is some partial service degradation with log drains configured to send to Mezmo where not all log messages are being sent. Mezmo is rejecting logs over a certain size. We are actively investigating this issue with the upstream provider.

2. 3 juin 2026 à 19:16 UTCresolved

   As of 18:42 UTC the upstream provider began to accept log messages again consistently and this issue is now remediated.

1. 8 mai 2026 à 05:02 UTCmonitoring

   Aptible is aware of “Dirty Frag,” a recently disclosed Linux kernel vulnerability that may allow local privilege escalation under certain conditions [0]. We have taken action across the Aptible platform to protect all customer workloads against this vulnerability. No customer action is required at this time. We will continue to monitor upstream kernel, distribution, and security vendor guidance related to this vulnerability and will take any additional action if needed. [0] https://www.openwall.com/lists/oss-security/2026/05/07/8

2. 8 mai 2026 à 15:35 UTCresolved

   This incident has been resolved.

1. 8 mai 2026 à 01:05 UTCinvestigating

   Due to an underlying issue in AWS we are experiencing instability in a certain availability zone in us-east-1. We are investigating this issue and will follow up with recommendations shortly

2. 8 mai 2026 à 01:39 UTCidentified

   We have blocked operations from provisioning into the specified availability zone

3. 8 mai 2026 à 14:02 UTCmonitoring

   We have unblocked operations from the specified region as we have seen successful provisioning and maintenance of new instance in that availability zone. Instances that were affected have not yet seen recovery, and the upstream AWS alert indicates they may still take hours of recovery.

4. 9 mai 2026 à 03:35 UTCresolved

   An upstream AWS issue affected infrastructure in a single Availability Zone in us-east-1. During the event, we blocked provisioning into the affected zone and monitored AWS’s recovery efforts. AWS has since recovered enough of the affected availability zone for the impacted Aptible infrastructure to recover. We have also completed recovery of any Aptible-managed resources that required follow-up action on our side. Any customers with directly affected resources have been notified of resolution. We are continuing normal platform monitoring, but we are marking this incident as resolved on our end.

1. 29 avril 2026 à 22:40 UTCresolved

   Aptible is aware of CVE-2026-31431, also known as “Copy Fail,” a recently disclosed Linux kernel vulnerability that may allow local privilege escalation under certain conditions [0]. We have taken action across the Aptible platform to protect all customer workloads against this vulnerability. No customer action is required at this time. We will continue to monitor upstream kernel, distribution, and security vendor guidance related to this CVE and will take any additional action if needed. [0] https://copy.fail/

1. 19 avril 2026 à 02:03 UTCinvestigating

   We are investigating an incident impacting api.aptible.com. Viewing current resources and running operations are currently impacted, however running applications and databases are unaffected

2. 19 avril 2026 à 02:54 UTCmonitoring

   A fix has been implemented and operations that were failing to enqueue during the availability outage have been retried. We will continue monitoring the situation.

3. 19 avril 2026 à 03:10 UTCresolved

   This incident has been resolved.

1. 17 mars 2026 à 22:33 UTCinvestigating

   We are currently observing intermittent DNS resolution failures affecting infrastructure in us-east-1. These resolver errors may cause some Aptible Deploy operations (such as deploys, restarts, SSH, or container provisioning) to fail or time out. Applications running on Aptible Deploy that rely on network hostname resolution may also experience intermittent connection failures related to DNS resolution. We are actively working with our infrastructure providers to investigate the issue and will provide updates here as they are available.

2. 17 mars 2026 à 23:02 UTCmonitoring

   We have observed DNS resolution error rates return to normal levels, and all dependent activities should be functioning normally. We will continue to monitor for an additional period to ensure the issue is fully resolved before closing the incident. Further updates will be posted here if we observe any recurrence.

3. 17 mars 2026 à 23:33 UTCresolved

   This incident has been resolved.

4. 18 mars 2026 à 16:42 UTCpostmortem

   Additional detail about timeline: we observed a relative increase in DNS resolver errors in us-east-1 beginning at approximately 2026-03-17 18:30 UTC. The elevated error rate was intermittent rather than constant, with the most visible clusters occurring from roughly 18:30-18:55 UTC and again from about 20:55-21:40 UTC. Error rates returned to normal shortly after 21:40 UTC.

1. 31 décembre 2025 à 10:48 UTCinvestigating

   We are aware of degraded performance affecting auth.aptible.com. This affects all actions that interact with our APIs including login, viewing and managing resources, etc.

2. 31 décembre 2025 à 11:33 UTCmonitoring

   Performance has improved since our mitigations. We are still monitoring the situation for continued degradation.

3. 31 décembre 2025 à 14:49 UTCresolved

   This incident has been resolved.

1. 30 décembre 2025 à 02:14 UTCmonitoring

   The Aptible Security Team is aware of the recently disclosed vulnerability CVE-2025-14847 (https://github.com/advisories/GHSA-4742-mr57-2r9j) affected all MongoDB versions. The vulnerability could allow an attacker with network access to a MongoDB database to exfiltrate data, including sensitive data and/or credentials, without authenticating to the database. In response to the vulnerability, we have updated our supported MongoDB versions to prevent exploitations of this vulnerability. In addition, we have proactively restarted databases matching either of the following criteria, to ensure they are running on the latest protected versions: * All databases in shared-tenancy stacks, and * All databases with customer-created public endpoints that do not have access restricted to an IP allow list Since Aptible databases run on private networks by default, most Aptible managed MongoDB databases are not accessible from the internet or by other Aptible customers, and so are not vulnerable to CVE-2025-14847. As such, we did not proactively restart these databases. Customers may restart their databases at any time to update to the latest protected versions.

2. 30 décembre 2025 à 17:15 UTCresolved

   This incident has been resolved.

1. 5 décembre 2025 à 14:39 UTCinvestigating

   We are aware of the recently disclosed critical vulnerability CVE-2025-55182 (https://react.dev/blog/2025/12/03/critical-security-vulnerability-in-react-server-components) affecting React Server Components, which could allow remote code execution under certain conditions. This vulnerability affects certain React 19.x and frameworks that implement React Server Components, including Next.js. *Aptible Platform Impact:* Aptible has reviewed all infrastructure components that we manage and has confirmed that all are unaffected by this vulnerability. *Customer Application Impact:* If you are running applications on Aptible that use React Server Components, you may be affected. We have seen active exploitation of this vulnerability, and we recommend upgrading to the patched versions immediately. Specifically, applications using: - React 19.0.0, 19.1.0, 19.1.1, or 19.2.0 with react-server-dom-webpack, react-server-dom-turbopack, or react-server-dom-parcel - Next.js versions 14.3.0-canary.77 and all subsequent 14.3.x canary releases, 15.0.0, 15.0.1, 15.0.2, 15.0.3, 15.0.4, 15.1.0, 15.1.1, 15.1.2, 15.1.3, 15.1.4, 15.1.5, 15.1.6, 15.1.7, 15.1.8, 15.2.0, 15.2.1, 15.2.2, 15.2.3, 15.2.4, 15.2.5, 15.3.0, 15.3.1, 15.3.2, 15.3.3, 15.3.4, 15.3.5, 15.4.0, 15.4.1, 15.4.2, 15.4.3, 15.4.4, 15.4.5, 15.4.6, 15.4.7, 15.5.0, 15.5.1, 15.5.2, 15.5.3, 15.5.4, 15.5.5, 15.5.6, 15.6.0-canary.0 through 15.6.0-canary.57, 16.0.0, 16.0.1, 16.0.2, 16.0.3, 16.0.4, 16.0.5, 16.0.6 - Other frameworks implementing React Server Components including Vite, Parcel, React Router, RedwoodSDK, Waku *Additional Resources:* React Security Advisory: https://github.com/facebook/react/security/advisories/GHSA-fv66-9v8q-g76r Next.js Security Advisory: https://github.com/vercel/next.js/security/advisories/GHSA-9qr9-h5gf-34mp We will continue to update this incident page as more information becomes available.

2. 8 décembre 2025 à 16:52 UTCresolved

   This incident has been resolved.

1. 5 novembre 2025 à 15:59 UTCidentified

   We are investigating an issue causing degraded performance for the Aptible Platform API. As part of the mitigation, a temporary operation block across all regions is in place which prevents new operations from starting. This includes deployments, scaling actions, SSH sessions, and database tunnels. Running apps and databases are not affected. Next update We will provide an update by 11:30 ET or sooner.

2. 5 novembre 2025 à 16:05 UTCmonitoring

   A fix has been implemented and we are monitoring the results. The operation block has been lifted.

3. 5 novembre 2025 à 16:07 UTCresolved

   This incident has been resolved.

1. 29 octobre 2025 à 21:17 UTCinvestigating

   We are currently experiencing failures when provisioning new hosts across all regions. This impacts operations that require new capacity, including but not limited to: app deploys, scaling events, and new database creation. Existing running workloads are not affected. Our team is actively investigating and working to restore normal operation performance. We will provide an update by no later than 5:45pm EST.

2. 29 octobre 2025 à 21:33 UTCmonitoring

   A fix has been implemented and we are monitoring the results.

3. 29 octobre 2025 à 21:50 UTCresolved

   This incident has been resolved.

1. 28 octobre 2025 à 17:09 UTCinvestigating

   We are seeing significantly increased provisioning times in the AWS us-east-1 region. This is affecting operations that require new capacity, including deployments, scaling, and creation of new resources. These operations may time out and fail. Our team is actively investigating and working to mitigate the impact. We will provide our next update by 2 pm EST.

2. 28 octobre 2025 à 17:32 UTCmonitoring

   AWS has confirmed the issue with EC2 provisioning issues in us-east-1. Meanwhile, we have been able to implement a mitigation for affected customers. We will continue to leave this issue open and monitor its progress, and will provide any relevant updates as needed.

3. 28 octobre 2025 à 17:53 UTCidentified

   We are now seeing additional broader EC2 provisioning issues in the AWS us-east-1 region that are outside the scope of our mitigation. Additional delays/failures on operations that require new capacity in us-east-1 may still occur at this time. We will provide our next update by 2:30 pm EST.

4. 28 octobre 2025 à 18:35 UTCmonitoring

   We are not currently seeing provisioning issues in AWS us-east-1. We are modifying the incident status to Monitoring, and will provide any relevant updates as needed.

5. 28 octobre 2025 à 21:08 UTCresolved

   This incident has been resolved.

1. 20 octobre 2025 à 11:53 UTCidentified

   AWS is experiencing an issue with their APIs in us-east-1, and many services are impacted at this time. This may lead to failed operations. There is currently no impact to availability for apps and databases running on Aptible at this time.

2. 20 octobre 2025 à 13:56 UTCidentified

   We continue to see fall out from the AWS outage this morning. Operations that result in scaling capacity such as service scales, deployments, etc. may be impacted, resulting in failing operations. We recommend periodically retrying such operations.

3. 20 octobre 2025 à 15:16 UTCidentified

   We’re continuing to see widespread impact from the ongoing AWS outage in us-east-1. Most requests for new EC2 instances in this region are currently failing. As a result, any operations that require additional instance capacity, such as service scales, deployments, and other scaling-related actions, are likely to fail. This remains an evolving AWS issue, and we’ll provide updates as more information becomes available. At this time, apps and databases operating with existing capacity appear to be running normally without impact.

4. 20 octobre 2025 à 15:44 UTCidentified

   We are seeing continued widespread impact from the AWS outage in us-east-1. AWS has confirmed ongoing network and API failures within the EC2 internal network that are preventing new instance launches across multiple Availability Zones. To avoid any operation failures that could leave resources in a transitional state, we have implemented an operation block for all new operations in the us-east-1 region. Apps and databases already running on existing capacity continue to operate normally at this time. We are actively monitoring AWS recovery efforts and will lift the block once we confirm increased stability in the us-east-1 region. Further updates will be posted as they become available.

5. 20 octobre 2025 à 18:04 UTCidentified

   We are starting to see EC2 provisioning operations in us-east-1 succeed again as AWS recovery efforts progress. We’re continuing to monitor closely, and if this trend holds, we expect to lift the operational block in the region shortly. We will provide another update as soon as the block is lifted or if conditions change.

6. 20 octobre 2025 à 18:55 UTCmonitoring

   We have lifted the operation block in us-east-1 as we are now seeing a sustained increase in successful EC2-related operations. While some operations may still fail as AWS recovery continues, the overall success rate has improved significantly. We are observing that EC2-related actions are taking longer to complete than usual, so users may experience increased delays in operation completion times. Given this progress, new operations have been re-enabled. We’ll continue to monitor closely and will update the incident status once failure rates and operation times have fully stabilized.

7. 20 octobre 2025 à 20:03 UTCmonitoring

   Platform operations in us-east-1 are fully operational at this time. We continue to actively monitor EC2-related operations for any elevated error rates. We’ll keep this incident open while AWS’s own outage remains active and will close it once AWS has resolved their incident and we’ve confirmed continued stability across all operations in us-east-1.

8. 20 octobre 2025 à 23:04 UTCresolved

   AWS has marked the underlying issue in us-east-1 as resolved, and we’ve confirmed continued stability across all EC2-related operations. All Aptible operations are functioning normally, and this incident is now closed.

1. 20 octobre 2025 à 07:47 UTCidentified

   AWS is experiencing an issue with their APIs in us-east-1, and many services are impacted at this time. This may lead to failed operations, but there appears to be no impact to availability for apps and databases running on Aptible at this time.

2. 20 octobre 2025 à 08:31 UTCmonitoring

   Since the AWS issues seem serious, and to avoid any operation failures causing resources to get stuck in a transitional state, we are taking action to block all new operations on the platform at this time. We have identified no impact to acceibilty or preformance of running apps and database containers at this time.

3. 20 octobre 2025 à 09:30 UTCmonitoring

   AWS has indicated a fix has been applied, and we are re-enabling operations. We will continue to monitor for additional impact, as both Dockerhub and Quay.io were offline due to the AWS issue.

4. 20 octobre 2025 à 09:46 UTCresolved

   This incident has been resolved.

1. 12 octobre 2025 à 02:30 UTCinvestigating

   One of our internal queuing systems observed a failure that has affected several operations enqueued in us-east-1. Our Reliability Team is restoring the queueing system and investigating the root cause of the failure.

2. 12 octobre 2025 à 02:42 UTCmonitoring

   A fix has been implemented and we are monitoring the results.

3. 12 octobre 2025 à 03:40 UTCmonitoring

   No operation queueing issues have been observed since the last update. If you observed any operations that failed or were stuck in a "pending" state, please retry the operations and they should run without errors.

4. 12 octobre 2025 à 03:40 UTCresolved

   This incident has been resolved.

1. 10 octobre 2025 à 15:53 UTCresolved

   Aptible has reviewed and addressed the Redis vulnerability described in https://redis.io/blog/security-advisory-cve-2025-49844/ . We have confirmed that all internal Redis instances used in delivering the Aptible platform are secure and have no exploitation path related to this CVE. For Aptible customers who do not have their Redis databases publicly exposed, a path to exploitation is similarly unlikely. Additionally, for customers using Redis 6.2 databases on Aptible, a patched version will be available by Oct 10th at 5pm EST, and additional releases (7.2) are upcoming; see aptible.com/changelog for ongoing release updates. To ensure you are running the latest minor version when available, please run the cli command: aptible db:reload for each Redis database. For more information on using this command, visit https://www.aptible.com/docs/reference/aptible-cli/cli-commands/cli-db-reload . If you have any specific questions or concerns related to this CVE, please contact us .