It was a relatively quiet this week in Cloud Dependencies. Of the most notable outages, we saw that:
- Atlassian had a major outage for 1 hour 56 minutes on the 22nd in which some but not all of the users saw eros in the web US for Jira, Confluence, and other products. Based on the timing of the social media reports, it took Atlassian more than half an hour to update their status page. See Jira for more information.
- Azure Cosmos DB had issues throughout the week in Azure East US. See the report below for more information on these issues.
To learn more about the other notable outages we saw, read on.
Notable Metrist-Reported Downtime
AWS
- AWS RDS was unable to create new instances in AWS US EAst 2 for 54 minutes on the 16th.
- AWS SQS had extreme latency for writing messages in AWS US WEst 1 with response times spiking more than 700% from usual for 3 hours 24 minutes on the 14th.
Azure
- Azure Cosmos DB had issues throughout the week in Azure East US. The platform was
- Unable to create Cosmos DB accounts in Azure East US for periods of 20-60 minutes multiple times per day throughout the week (from the 16th to the 25th).
- General issues in Azure East US throughout the day on the 19th and 20th.
- Unable to create databases for 19 minutes on the 23rd.
- Unable to delete databases for 19 minutes on the 21st. The issue was also seen in Azure Canada Central for 20 minutes on the same day.
- Unable to create containers for 29 minutes on the 23rd and 20 minutes on the 27th.
- Azure AKS was unable to create deployments in Azure Central US for 30 minutes on the 18th. On the 21st it was unable to create clusters in Azure West US 2 for 1 hour 7 minutes.
- Azure VMs was unable to create new instances in Azure West US for 28 minutes on the 19th.
- Azure App Service’s Ping Service was not responding from Azure Central US and was down for 59 minutes on the 25th.
GCP
- GCP Compute Engine was unable to delete instances inGCP US Central 1 for 41 minutes on the 21st.
Saas/Other
- Circle CI was unable to run Docker workflows for 20 minutes on the 14th.
- PagerDuty was unable to send webhooks for 1 hour 46 minutes on the 14th. Metrist saw warning signs about 1 hour prior to the issue starting.
- Hubspot was unable to get contacts from all tested regions for 9 minutes on the 22nd.
- Jira was unable to create issues in US West 1 and was down for 12 minutes on the 24th.
- Zoom was unable to join calls from US East 2 for four separate 8-16 minute intervals throughout the course of the day on the 24th.
