Major incidents SaaS vendor changes, page 11

Tailscale reported a major Funnel incident involving connection errors for some users, starting 2026-09-01 12:47 UTC and resolved at 2026-09-01 16:19 UTC, lasting 3 h 31 min. A fix was deployed at 09-01 14:11 UTC; no cause was stated.

Dynatrace reported a major authentication issue affecting service access and potentially causing data processing and extraction errors or delays; ingestion continued. The incident started at 2026-09-01 11:54 UTC and was resolved at 2026-09-03 14:10 UTC, lasting 2 days 2 h, with normal operation verified after extended monitoring.

Snowflake reported a major incident in which customers in the listed AWS, Azure and GCP regions could be unable to sign in to Snowflake Cowork. It started at 2026-09-01 09:58 UTC and was resolved at 2026-09-01 13:50 UTC. All other Snowflake features were unaffected; Snowflake identified the source but did not specify the cause or fix.

A critical Supabase Dashboard outage prevented login, logged existing users out, and made Dashboard actions unavailable from 2026-09-01 08:45 UTC to 2026-09-01 10:34 UTC (1 h 49 min). Supabase reverted a recent change that affected login; the incident was resolved, and projects and non-Dashboard traffic remained unaffected.

Elastic Cloud Hosted reported major degradation from 2026-09-01 08:16 UTC to 2026-09-30 20:44 UTC (29 days 12 h). In Elasticsearch 9.5.0, 9.5.1 and 9.5.2, bulk indexing could cause replica divergence, high disk usage and long-running recoveries. Elastic fixed the bug in 9.5.3 and advised affected customers to upgrade and follow workaround steps.

Sentry reported a major outage affecting API, Dashboard, and Integration Pipeline, with requests timing out or returning 504s. It started at 2026-09-01 08:01 UTC and was resolved at 2026-09-01 11:21 UTC, lasting 3 h 20 min. Sentry identified the issue and rolled out a configuration change.

Salesforce reported a major Marketing Cloud Core Service disruption affecting 1 instance. It started at 2026-09-01 00:27 UTC and was resolved at 2026-09-01 00:38 UTC, lasting 11 min.

Discord reported a major outage that left some users unable to connect to some servers and affected voice calls and activities. It started at 2026-08-31 22:54 UTC and was resolved at 2026-09-01 00:49 UTC, lasting 1 h 55 min; Discord reported remediation steps but gave no specific cause or fix.

Rippling reported a major outage preventing some saved payroll reports from loading when certain payroll fields were included. It started at 2026-08-31 22:37 UTC and lasted 5 h 43 min, resolving at 2026-09-01 04:20 UTC; United States access was restored before European Union access.

Jira reported a major outage fetching mail from Microsoft accounts, disrupting email-based creation and updates of work items and potentially returning 500 errors. It started at 2026-08-31 18:28 UTC and was resolved at 2026-09-01 05:18 UTC, lasting 10 h 49 min; Atlassian reported full recovery across impacted services.

Ashby reported a major degradation affecting All-in-One document uploads and signing through Dropbox Sign, starting 2026-08-31 17:44 UTC and resolved at 2026-08-31 18:11 UTC (27 min). Ashby attributed the disruption to an upstream Dropbox Sign incident and said the provider resolved the issue.

Snowflake reported a critical outage in AWS - US East (N. Virginia) affecting sign-in, queries and data operations from 2026-08-31 17:20 UTC to 2026-08-31 19:44 UTC (2 h 24 min). It reverted a problematic change and fixed residual Snowsight access and Snowpipe Streaming impact; an RCA is available on Snowflake Community.

Render reported a major degradation in Singapore affecting deploys and inbound traffic, with unexpected web service restarts or intermittent availability; Free service spin-ups from idle were temporarily disabled. The incident started at 2026-08-31 15:47 UTC and was resolved at 2026-08-31 18:09 UTC, lasting 2 h 21 min. Render implemented a fix but did not specify its details.

Asana reported a major full outage for some users (25% outage), affecting API, App and Mobile, including US components, from 2026-08-31 15:14 UTC to 2026-08-31 17:18 UTC (2 h 4 min). The postmortem attributed it to infrastructure changes and peak traffic exceeding cache scaling limits; user-facing symptoms recovered after configuration changes and a fix.

ChatGPT Work had a major incident with elevated errors and latency, preventing users across multiple subscription plans from starting or continuing tasks; Plus users were particularly impacted, with Work mode unavailable. It started at 2026-08-31 15:04 UTC and resolved at 2026-08-31 20:28 UTC, lasting 5 h 23 min. A mitigation was applied, and services fully recovered.

ChatGPT Work had a major incident starting 2026-08-31 15:04 UTC, lasting 5 h 23 min and resolving at 2026-08-31 20:28 UTC. Elevated errors and latency prevented users across multiple subscription plans from starting or continuing tasks; Plus users were particularly impacted, with Work mode unavailable. OpenAI applied a mitigation and reported full recovery.

Miro reported a major incident causing intermittent registration and login errors for a small percentage of users, starting 2026-08-31 14:31 UTC and resolved at 2026-08-31 15:57 UTC, lasting 1 h 25 min. Miro implemented a fix and confirmed full recovery; users encountering registration errors were advised to refresh and retry.

Baseten reported a major Dedicated Inference incident with elevated inference 500s for models in a single US-based cluster. It ran from 2026-08-31 06:38 UTC to 2026-08-31 07:02 UTC (24 min); a fix was in place and traffic was back at 06:57 UTC.

Xero reported a major Australian Xero Tax outage preventing some customers from prefilling or lodging tax returns to the ATO, starting 2026-08-30 23:50 UTC and resolved 2026-08-31 01:37 UTC (computed duration: 1 h 46 min). Xero attributed the issue to an underlying product dependency and implemented a fix to restore service.

Xero reported a major incident preventing Australian Xero Tax customers from prefilling or lodging tax returns to the ATO. It ran from 2026-08-30 20:21 UTC to 2026-08-30 22:39 UTC, lasting 2 h 18 min. The ATO implemented a fix, and Xero confirmed resolution.

Some Notion AI MCP tool calls failed in a major incident that started at 2026-08-30 16:13 UTC and was resolved at 2026-08-30 22:49 UTC, lasting 6 h 36 min. Notion said engineering had identified the issue and was working to mitigate it, but provided no specific cause or fix.

Salesforce reported a major Core Service outage blocking B2B Commerce on Core cart/checkout workflows in most sandbox environments because underlying API calls failed. Recorded from 2026-08-29 03:07 UTC to 2026-08-31 02:40 UTC (47 h 33 min), it was marked resolved; Salesforce validated a fix-forward and began a staggered rollout.

Upstash reported a major QStash message-persistence disruption in us-east-1 and said a fix restored service. The listed start is 2026-08-28 22:56 UTC, but resolution is recorded as 2026-08-28 22:20 UTC and the update describes disruptions between 22:20 and 22:33 UTC; duration is unknown.

Otter.ai reported a major outage affecting notification, invite, password reset and payment emails, reminders, and bulk export. It started at 2026-08-28 19:09 UTC and was resolved at 2026-08-28 20:18 UTC, lasting 1 h 8 min; Otter stated that systems were fully restored.

ActiveCampaign reported a major incident affecting access to some accounts, listing Platform Availability and US Region - Platform Availability. It started at 2026-08-28 18:48 UTC and was resolved at 2026-08-28 19:16 UTC; the reported duration was 27 min. A fix was implemented at 19:03 UTC; no cause was stated.

Anthropic reported a major outage affecting Claude Cowork and Claude Code on the web: some sessions failed to start or disconnected mid-task due to an upstream cloud provider issue. It started at 2026-08-28 17:22 UTC and was resolved at 2026-08-28 20:21 UTC, lasting 2 h 59 min. A mitigation was applied; affected sessions could be retried.

Claude reported a major incident affecting Claude Cowork and Claude Code on the web from 2026-08-28 17:22 UTC to 2026-08-28 20:21 UTC, lasting 2 h 59 min. Some sessions failed to start or disconnected mid-task due to an upstream cloud provider issue; affected sessions could be retried. A mitigation was applied before resolution.

Grafana Cloud reported a major partial Logs write outage in AWS US East - prod-us-east-3, also affecting Frontend Observability in that region. It started at 2026-08-28 17:12 UTC and was resolved at 2026-08-28 18:42 UTC, lasting 1 h 29 min. Grafana reported deploying a fix and stabilizing service.

Bitrise reported a major incident affecting Builds (CI): the Builds List page failed to load in multiple workspaces. It started at 2026-08-28 13:01 UTC and was resolved at 2026-08-28 13:33 UTC, lasting 32 min; no cause or fix was stated.

Dynatrace reported a major Support Portal outage: customers encountered errors communicating through support cases, while chat remained available. The incident started at 2026-08-28 12:53 UTC and was resolved at 2026-08-28 16:30 UTC, lasting 3 h 37 min; Dynatrace confirmed corrective actions restored case comments and communication.

Box reported a major-impact incident affecting access to Box Hubs through shared links, starting 2026-08-28 08:20 UTC and resolved at 2026-08-28 10:00 UTC (1 h 40 min). Box said it identified the underlying cause and restored full functionality, but the evidence does not specify the cause or fix.

Mailgun reported a critical EU datacenter messaging outage affecting API, SMTP, inbound processing, outbound delivery, events and logs, analytics, and the control panel. It started at 2026-08-27 20:38 UTC and resolved at 2026-08-27 21:16 UTC, lasting 37 min. Mailgun implemented a fix but did not specify the cause.

Cursor reported a critical sign-in service degradation affecting cursor.com and IDE from 2026-08-27 19:31 UTC to 2026-08-27 19:38 UTC, lasting 7 min. A fix was implemented at 19:35 UTC, and Cursor confirmed sign-in had returned to normal at resolution.

Snowflake reported a major outage affecting private connectivity, including PrivateLink, in Azure - Australia East (New South Wales), Azure - Central India (Pune), and Azure - Central US (Iowa), causing connection errors or timeouts. It ran from 2026-08-27 19:05 UTC to 2026-08-27 20:26 UTC (1 h 20 min); Snowflake implemented a fix and confirmed service restoration.

Dynatrace reported major intermittent performance degradation affecting some customer environments in the US East region, starting 2026-08-27 18:18 UTC. The issue was mitigated by 08-27 18:49 UTC and resolved at 2026-08-31 08:50 UTC after extended monitoring; the recorded incident duration was 3 days 14 h. No specific cause was disclosed.

Some Serverless customers’ .apm endpoints were unavailable after a Managed Inputs change removed internal DNS rules. The major incident was recorded from 2026-08-27 17:11 UTC to 2026-08-27 18:33 UTC (1 h 22 min), resolved. The update instead described disruption within a 9 hour window on August 26, with service restored around 9AM UTC that day.

Gorgias reported a major incident causing ticket, order and product searches to show outdated results in the Helpdesk Web App and us-east1-c94b. It started at 2026-08-27 16:22 UTC and was resolved at 2026-08-27 17:02 UTC, with a computed duration of 39 min; a fix was implemented at 16:43 UTC.

Omni Analytics reported a critical incident affecting Ireland: server timeouts caused 500 errors on EU instances. It started at 2026-08-27 14:52 UTC and was resolved at 2026-08-27 15:47 UTC, lasting 55 min; the vendor reported that one model service spiked and began recovering.

Smartsheet reported a critical outage affecting login, sheet access, Control Center, Dynamic View and other premium apps, disrupting access to customer work. It started at 2026-08-27 13:33 UTC and resolved at 2026-08-27 15:53 UTC, lasting 2 h 19 min; a fix was implemented at 08-27 13:52 UTC.

Opsgenie reported a major Microsoft Teams notification outage for standalone customers, starting 2026-08-27 10:26 UTC and resolved 2026-09-03 06:06 UTC (6 days 19 h); Jira Service Management Operations was unaffected. An external partner’s operational error caused the issue; partner corrective measures restored service, though customers still affected were directed to restore their integration.

GitHub reported a critical degradation affecting Kimi K3 in Copilot products and IDE surfaces, with elevated error rates; no other models were impacted. The incident record started at 2026-08-27 10:04 UTC and resolved at 2026-08-27 12:12 UTC, lasting 2 h 8 min. GitHub attributed it to an upstream provider issue, resolved through the provider’s mitigation.

GitHub reported a critical degradation affecting Kimi K3 in Copilot products and IDE surfaces, with elevated error rates; no other models were impacted. The incident started at 2026-08-27 10:04 UTC and was resolved at 2026-08-27 12:12 UTC, lasting 2 h 8 min. GitHub attributed it to an upstream provider issue, resolved through the provider’s mitigation.

Sentry reported a major outage affecting product and API availability, with the Dashboard and US ingestion components listed as affected. It started at 2026-08-27 09:24 UTC and was resolved at 2026-08-27 15:41 UTC, lasting 6 h 17 min; latency returned to normal.

A critical outage made Confluence, Jira Cloud and Jira Service Management unavailable to affected users in Western Europe and Southeast APAC. The incident ran from 2026-08-27 04:22 UTC to 2026-08-27 05:16 UTC (53 min); Atlassian rolled out a fix and confirmed recovery. The postmortem attributed the disruption to a defective routing-layer code change in a proxy fleet.

A critical outage made Jira Cloud, Jira Service Management and Confluence unavailable to affected users in Western Europe and Southeast APAC. The incident ran from 2026-08-27 04:22 UTC to 2026-08-27 05:16 UTC (53 min); Atlassian rolled out a fix, and its postmortem attributed the disruption to a defective proxy routing-layer code change.

Klaviyo reported a major degradation starting 2026-08-27 00:09 UTC: some profiles stalled at flow time delays, affecting email, text messaging, push, and WhatsApp sending. Campaigns and flows without time delays were unaffected. A fix was deployed and affected-message processing completed; the incident resolved at 2026-08-27 03:42 UTC after 3 h 33 min.

Some Buffer customers could not log in because email verification codes failed to arrive or were rejected. The vendor-rated major outage started at 2026-08-26 23:29 UTC and was resolved at 2026-08-27 03:48 UTC, lasting 4 h 18 min. Buffer applied a fix restoring verification email delivery and working codes.

Apollo GraphQL reported a major incident starting 2026-08-26 22:12 UTC: user invites, user permission changes and OAuth changes intermittently failed because rate limits were incorrectly applied. It lasted 1 h 24 min and was resolved at 2026-08-26 23:36 UTC after a fix was deployed to ensure rate limits work as expected.