58 Cloudamqp incidents · décembre 2024 — official updates, affected components, duration and resolution details.
We're investigating connectivity issues affecting some servers in Azure West Europe. We'll post an update as soon as we know more.
Connectivity to servers in southeast Asia has been restored.
Nous enquêtons sur les problèmes de connectivité touchant certains serveurs de Google europe-west2. Nous publierons une mise à jour dès que nous en saurons plus.
Nous continuons d'enquêter sur cette question.
Incident est limité aux serveurs en europe-west2-b
GCE continue de signaler des problèmes, mais nous voyons certains appels d'API réussir.
GCE a confirmé qu'une panne de courant en Europe-Ouest2-b a causé la perturbation. La question de la puissance a été atténuée, et le recouvrement complet des charges de travail affectées est provisoirement attendu d'ici 13h UTC aujourd'hui. Nous continuons de surveiller le rétablissement.
Nous restons à l'affût de tout autre problème.
Cet incident a été résolu.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
Nous enquêtons sur des problèmes de connectivité touchant certains serveurs d'Azure Ouest Europe. Nous publierons une mise à jour dès que nous en saurons plus.
Azure est au courant de cette question et enquête activement. Nous publierons une autre mise à jour dès que nous en saurons plus.
Azure a identifié ce problème. Ils ont réduit la question à l'accessibilité du stockage qui touche la région. Leurs équipes mènent activement des actions de rétablissement sur l'infrastructure touchée tout en continuant d'enquêter sur la cause sous-jacente. Nous suivons les progrès de la reprise et nous travaillons à rétablir les ressources affectées.
Nous avons été en mesure d'établir une communication avec les instances d'Azure en Europe occidentale. La majorité des instances sont en marche. Azure a encore des MV à résoudre.
Toutes les instances CloudAMQP sont en cours d'exécution à Azure West Europe.
** Résumé** CloudAMQP a perdu la connectivité avec un sous-ensemble de clusters dans la zone de disponibilité 2 d'Azure West Europe le 9 octobre. Les grappes de nœuds uniques dans l'AZ2 ont été les plus touchées par l'incident. (en milliers de dollars) **Impact** Les clients n'ont pas été en mesure d'accéder aux données ou aux nœuds de mange touchés pendant l'incident. Chronologie des événements d'Azure. * 12:25 UTC le 09 Octobre 2026 – Nous avons identifié ce problème via nos alertes de surveillance du service interne. * 12:29 UTC le 09 Octobre 2026 – L'impact client a commencé, avec des défaillances de disque VM détectées dans l'unité de stockage de l'Europe occidentale touchée. * 2h30 UTC le 09 octobre 2026 – L'enquête a identifié la perte de battements cardiaques dans plusieurs nœuds de stockage et domaines de défaillance, ce qui a conduit à l'unité d'échelle de stockage pour aller en mode de détresse. * 15:32 UTC le 09 Octobre 2026 - Nous avons pris l'échelle de stockage hors du mode de détresse, isolant les disques sur les serveurs de stockage touchés pour permettre la récupération des disques restants. À ce stade, la plupart des clients touchés ont commencé à se rétablir. * 21:00 UTC le 09 Octobre 2026 - Datacenter et les ingénieurs de stockage ont continué à exécuter le plan de récupération sur plusieurs nœuds de stockage. * 23:27 UTC le 09 Octobre 2026 – Les efforts pour récupérer les nœuds de stockage touchés ont été couronnés de succès, et les travaux pour récupérer les disques virtuels ont continué. Les mesures de rétablissement subséquentes comprenaient le rétablissement au niveau des nœuds et le rétablissement des ressources de stockage touchées. * 03:00 UTC le 10 octobre 2026 – La fonctionnalité du service a été récupérée, et l'impact client a été atténué. ** Cause de la mort** Azure a connu un problème de plateforme a eu un impact sur les disques de stockage Azure Virtual Machines dans la région de l'Europe de l'Ouest AZ 2. Les clients touchés peuvent avoir subi des défaillances de connectivité, des redémarrages inattendus ou des Machines virtuelles entrant dans un état défaillant dans la région. (en milliers de dollars) **Résolution** L'accès à la machine a été restauré par Azure en environ 3 heures. CloudAMQP a pu restaurer la communication avec les nœuds touchés. Nous avons continué à surveiller les mises à jour d'Azure pour nous assurer que les nœuds n'étaient plus touchés.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
Nous assistons à une détérioration de la connectivité réseau entre certains réseaux non Azure et Azure UAE North (azure-arm::uaenorth), causée par un problème de routage chez un fournisseur de transport en amont. Tous les serveurs de la région sont en place et traitent les messages, et la surveillance, les métriques et les journaux fonctionnent normalement. Les clients dont les clients se connectent à partir de réseaux touchés par ce problème de routage peuvent également voir des délais de connexion. Nous avons signalé le problème au fournisseur et acheminons notre propre trafic de gestion autour.
Le routage est de retour à la normale.
Cet incident a été résolu.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
Nous enquêtons actuellement sur cette question.
Le problème a été identifié et une solution est en cours de mise en oeuvre.
Un correctif a été mis en place et nous suivons les résultats.
Cet incident a été résolu.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
DigitalOcean connaît actuellement des problèmes avec son API et son panneau de contrôle et vous pouvez rencontrer des erreurs et des latences plus élevées pendant que le problème persiste
Cet incident a été résolu.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
Nous avons reçu diverses alertes nous informant que deux serveurs déployés sur Azure West US n'étaient pas accessibles. Apparemment, cela semble être lié spécifique à cette région impactant 25% des serveurs.
Microsoft Azure a confirmé qu'ils avaient des problèmes de réseau sur la région ouest des États-Unis. Pour l'instant, la plupart des serveurs sont opérationnels. Nous continuerons de surveiller les heures suivantes.
Cet incident a été résolu.
Traduit automatiquement depuis la mise à jour officielle de l'incident.
We are currently investigating this issue.
The issue has been identified and a fix is being implemented.
A fix has been implemented and we are monitoring the results.
This incident has been resolved.
**Summary** Between June 25 and June 26, our monitoring system incorrectly reported some dedicated servers as unreachable and sent "server down" notifications for servers that were in fact operating normally. This was a monitoring-side issue only — no customer servers or services were actually down, and no data or message delivery was affected. **Impact** Affected customers received one or more alerts indicating their dedicated server was down. These alerts were false positives. The servers themselves remained healthy and fully available throughout the incident. **Root cause** Before connecting to a server, our monitoring system performs a network authorization step. A defect in how one component handled a transient network hiccup could leave a monitoring process unable to complete that step, after which it failed to connect to the servers it was responsible for checking. The servers themselves stayed healthy and available throughout — the monitoring process had simply lost its ability to reach them, and reported them as down. **Resolution** We identified the affected monitoring processes, confirmed the root cause in production, and deployed a fix that makes this authorization step resilient to such transient failures and able to recover automatically. After deploying, we monitored the system to confirm that notifications returned to normal. We apologize for the confusion these false notifications may have caused.
Between 17:30 and 20:00 UTC on 2026-05-30, our server monitoring system generated a large number of false "server_unreachable" alerts. Clusters were operational and healthy throughout this period — no customer data or message delivery was affected. The alerts were caused by a rolling restart of our monitoring service, during which individual monitoring workers went offline mid-cycle and lost state. This caused the workers to report connection timeouts for servers they could no longer reach — even though those servers were fully healthy. We apologize for any concern this may have caused.
This incident has been resolved.
We are investigating connectivity issues affecting some CloudAMQP instances in Azure West Europe. Affected instances may see connection failures or unexpected restarts. The root cause is an ongoing Microsoft Azure platform incident in the West Europe region (started 14:31 UTC, 23 May 2026). Microsoft is investigating.
Resolved — Microsoft has confirmed that the Azure West Europe Virtual Machines incident is resolved. The impact window was 14:09–14:13 UTC on 23 May 2026, during which a limited number of customers may have experienced connection failures or unexpected Virtual Machine restarts. The Azure environment self-healed and the service has been confirmed restored. Affected CloudAMQP instances should now be operating normally.
Since around 8pm UTC last night, some metrics are not being forwarded to legacy integration. Prometheus integrations not affected. We are investingating.
A fix has been implemented and we are monitoring the results.
This incident has been resolved.
# Legacy host metrics integrations degraded Incident window: 2026-05-21 18:37 UTC – 2026-05-22 04:05 UTC \(9h 28m\) Affected service: Legacy metric integrations \(CloudWatch, Datadog, Librato, New Relic, Splunk, Stackdriver\). Host metrics affected, not broker metrics, nor was internal monitoring or console graphs. ## Summary For ~9.5 hours, the service that collects host-level metrics \(CPU, memory, disk, network\) for legacy third-party integrations entered a crash loop and stopped shipping data. ## Root Cause An internal credential-signing service was migrated to a new container runtime the afternoon before. Due to a bug with environment variables the collector could not renew it's credentials. The collector's existing credential remained valid for several hours, so the failure only surfaced when it tried to renew. ## Resolution On-call staff detected the failure early morning, developers helped restore the service and collector recovered at 04:05 UTC. ## Prevention * All services, both signing and collector report failed authentication attempts earlier and with higher severity * Metrics pipeline alarm thresholds was tightened so a similar drop in data triggers alarms faster.
We are currently experiencing issues with our backend services handling account and server creation. This does not affect any running customer servers but might delay provisioning new servers.
The issue has been identified and a fix is being implemented.
A fix has been implemented and we are monitoring the results.
This incident has been resolved.
We are currently experiencing issues with our backend services handling account and server creation. This does not affect any running customer servers but might delay provisioning new servers.
A fix has been implemented and we are monitoring the results.
Still seeing slow requests, doing another investigation.
Looks like we're back at full performance now.
This incident has been resolved.
We are currently investigating this issue.
Servers in Azure West Europe can experience connection issues due to underlying issues at provider. Latest update from Azure "Current Status: We continue to investigate an availability issue caused by a subset of unhealthy backend storage infrastructure supporting Azure Virtual Machines and Azure Direct Drive in West Europe."
This incident has been resolved.
We are currently experiencing issues with our backend services handling account and server creation. This does not affect any running customer servers but might delay provisioning new servers.
The issue has been identified and a fix is being implemented.
A fix has been implemented and we are monitoring the results.
We had an abnormal increase in traffic causing timeouts and slow responses for customer.cloudamqp.com, we increased capacity and are back to normal. This did not affect any running clusters, just provisioning and configuration updates.
We are getting indications of instances being down in Azure region westus. We are investigating the issue.
This incident has been resolved.
Servers in Azure West US 3 can experience connection issues due to underlying issues at provider. We monitor and will update when we know more.
This incident has been resolved.
There appears to be an issue regarding sending metrics to all Legacy metrics integrations. Alarms based on broker metrics also got affected. We are currently investigating the issue Prometheus based metrics (Datadog v3 etc) not affected.
A fix has been implemented and we are monitoring the results.
This incident has been resolved.
Broker metrics and metrics integrations degradation Window: 2026-04-23 03:25 – 05:35 UTC \(2h 10m\) ## Impact Broker metrics \(connections, channels, queues, consumers, message rates, node and netsplit state\) were degraded, progressively dropping until recovery at 05:35 UTC. Downstream effects: * Queue, Consumer, Connection, Channel, Connection flow, and Netsplit alarms: threshold breaches during the window may not have triggered, or triggered late. Alarms already tripped before 03:25 UTC kept their state. * Metrics Integrations \(Datadog, CloudWatch, New Relic, Splunk, Dynatrace, etc.\): delivery of broker metrics was reduced for the duration of the window. Alarms configured on the receiving side may likewise have missed or lagged. Not affected: server metrics \(CPU, memory, disk\) and their corresponding alarms, Notice alarms, broker availability, and the metric graphs shown in the CloudAMQP console \(these use a separate data path and continued to update normally\). ## Timeline \(UTC\) * 03:25 — broker-metrics collection begins degrading * 03:35 — ~30% of broker-metric samples missing * 03:50 — ~50% plateau * 05:25 — collector workers cycled and re-initialised * 05:35 — full throughput restored ## Root cause Broker metrics are polled from each cluster's management HTTP API by a pool of collector workers and published to an internal message bus that both the Alarms service and Metrics Integrations consume from. At 03:25 UTC this service became unresponsive without raising an error. Affected workers silently stopped polling while remaining alive to the platform, so automatic restart did not trigger. On-call staff began investigating around 05:00 UTC; force-restarting the service restored metric flow. The exact trigger has not been identified. Our focus is on ensuring the condition is detected and handled promptly if it recurs. ## What we are changing * Added metrics and alarms specific to this service, with high-urgency paging for on-call. * Reliability improvements in the service itself to detect and restart stalled work. ## What you can do CloudAMQP offers a new generation of metrics integrations based on Prometheus, with a re-engineered pipeline that does not rely on these centralised services — each server forwards data directly to your endpoint. These have been running in production for some time and have proved very reliable. * [https://www.cloudamqp.com/blog/prometheus-metrics-integrations.html](https://www.cloudamqp.com/blog/prometheus-metrics-integrations.html) * [https://www.cloudamqp.com/docs/monitoring\_metrics\_datadog\_v3.html](https://www.cloudamqp.com/docs/monitoring_metrics_datadog_v3.html) * Background: [https://www.cloudamqp.com/blog/decentralized-observability-with-open-telemetry-part-1.html](https://www.cloudamqp.com/blog/decentralized-observability-with-open-telemetry-part-1.html)
One of our backend components suffered a small outage due to an unexpected full services restart. Connectivity with the servers were broken and service triggered alarms for many clusters. We apologize for the inconvenience. All clusters were operational and healthy within the incident/alarms timespan.
Customers are experiencing Management UI being unavailable with 502 Bad Gateway. Timeline: - 14:16 UTC -> The issue was identified - 14:30 UTC -> Recent code changes were reverted. No relationship with incident noted.
The Management UI and brokers are serving traffic normally. You can still use the Management UI by using the Broker URL and properly typing User and Password. The impact is related to our backend service logs into Management UI via CloudAMQP SSO.
We are continuing to work on a fix for this issue.
The Management UI SSO shortcut is back to service. Impacted Clusters: All customers with clusters created more than 6 months ago Cause: Some code changes unexpectedly changed the communication schema between our backend services and the SSO component on the clusters. The proxy configuration conflicted with those changes causing the 502 Bad Gateway error to show up. Impact: Only the SSO via CloudAMQP Console was impacted given it required communication with our backend components. The AMQP Clusters were healthy and servicing traffic via both AMQP and HTTP on Management UI and API. Logging into the Management UI via Credentials and all Authentications Backends was functioning normally.
