Major incidents SaaS vendor changes, page 10

Bunny Stream had a major playback and upload outage from 2026-09-03 13:47 UTC to 2026-09-03 15:38 UTC (1 h 51 min), caused by a backend database issue triggering video-server timeouts. The issue was resolved; failed uploads need re-uploading, while existing transcoding videos were retained but could show inaccurate status.

Cursor reported a critical service degradation affecting All Grok Models, Automations, Cloud Agents, Grok Bot and Review Agents; IDE was also listed as an affected component. It lasted 3 h 26 min, from 2026-09-03 13:41 UTC to 2026-09-03 17:07 UTC. Cursor reported identifying the cause and working on mitigation, but disclosed no cause or fix details.

Anthropic reported a major incident causing elevated request errors across claude.ai, Claude API (api.anthropic.com), Claude Code, and Claude Cowork, initially affecting Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5. Starting 2026-09-03 13:26 UTC, it was resolved at 2026-09-03 16:23 UTC (2 h 57 min); a fix was deployed.

Claude reported a major incident with elevated errors affecting multiple models across claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork. It started at 2026-09-03 13:26 UTC and was resolved at 2026-09-03 16:23 UTC, lasting 2 h 57 min. A fix was deployed; the vendor said customer impact ended at 16:16 UTC.

Claude Sonnet 5 requests experienced elevated errors across claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork in a major-impact incident starting 2026-09-03 12:37 UTC. A fix was implemented by 12:47 UTC; the incident was resolved at 2026-09-03 12:56 UTC, with a reported duration of 18 min.

Claude Sonnet 5 requests experienced elevated errors across claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork, with major impact. The incident started at 2026-09-03 12:37 UTC and was resolved at 2026-09-03 12:56 UTC; the reported duration was 18 min. Anthropic implemented a fix and monitored results before resolution.

Strapi Cloud reported a critical outage of dashboard backup features starting 2026-09-03 12:16 UTC, lasting 1 h 26 min and resolving at 2026-09-03 13:42 UTC. A backup provider issue caused the unavailability; Strapi said backups were safe, nothing was deleted, and backups appeared as expected after resolution.

Grafana Cloud reported a major incident in AWS India - prod-ap-south-1, with increased latency and ingestion issues affecting deployments. It started at 2026-09-03 09:32 UTC and was resolved at 2026-09-03 12:49 UTC, lasting 3 h 17 min. A fix was applied, with recovery reported at 09-03 11:37 UTC.

Salesforce reported a major outage of Storefront and Business Manager for a subset of B2C Commerce Cloud customers across 3 instances, from 2026-09-03 06:43 UTC to 2026-09-03 12:36 UTC (5 h 53 min). A misconfigured legacy setting disrupted application connectivity following a release; configuration corrections and application restarts restored access.

ChatGPT Work Mode experienced high error rates in a major-impact incident starting at 2026-09-03 00:04 UTC. It lasted 6 min and was resolved at 2026-09-03 00:10 UTC, when all impacted services had fully recovered.

ChatGPT Work Mode experienced high error rates, with vendor-rated major impact, starting at 2026-09-03 00:04 UTC. The incident lasted 6 min and was resolved at 2026-09-03 00:10 UTC, when OpenAI reported all impacted services had fully recovered.

Discord reported a major voice endpoint selection outage starting 2026-09-02 23:26 UTC; calls could fail to start or get an endpoint. Discord identified an issue with its endpoint selection service and applied a mitigation, resolving the incident at 2026-09-02 23:42 UTC after 16 min.

Fly.io reported a major Sprites API partial outage affecting a subset of Sprites users, starting 2026-09-02 23:03 UTC and resolved at 2026-09-03 00:41 UTC, lasting 1 h 37 min. Error rates decreased during monitoring; no cause or fix was stated.

Claude reported a major incident involving elevated Claude Sonnet 5 request errors across claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork. The incident started at 2026-09-02 21:17 UTC and was resolved at 2026-09-02 21:44 UTC (computed duration: 26 min); the vendor stated actual impact occurred from 21:05 UTC to 21:19 UTC.

Claude Sonnet 5 requests experienced elevated errors affecting claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork. Anthropic rated the incident major; it started at 2026-09-02 21:17 UTC and was resolved after 26 min. The resolution update separately reported customer impact from 21:05 UTC to 21:19 UTC.

Asana reported a major outage affecting a subset of customers across API, App, Mobile and US components from 2026-09-02 18:08 UTC to 2026-09-02 18:49 UTC (41 min). A deployment rollback restored access; the postmortem attributed failed requests to insufficient cluster capacity after an automatic scaling system stopped replacing capacity removed during maintenance.

Render reported major builds/deploys degradation affecting Frankfurt, Ohio, Oregon, Singapore and Virginia, starting 2026-09-02 17:19 UTC. A fix was implemented by 2026-09-02 17:30 UTC, and the incident was resolved at 2026-09-02 17:40 UTC; the recorded duration was 20 min.

WorkOS reported a major AuthKit and Core Services - AuthKit outage affecting static asset delivery to end users with strict network allowlists. It started at 2026-09-02 16:39 UTC and was resolved at 2026-09-02 19:06 UTC, lasting 2 h 26 min. WorkOS mitigated the issue and confirmed delivery rates had recovered.

PlanetScale reported a major incident affecting postgres database management pages and API endpoints, which returned 500 errors. The incident started at 2026-09-02 16:17 UTC and was resolved at 2026-09-02 16:44 UTC, lasting 27 min; PlanetScale deployed a fix.

Some Discord servers were offline or degraded during a major-impact incident starting 2026-09-02 16:12 UTC and resolved at 2026-09-02 18:26 UTC, lasting 2 h 14 min. Discord reported recovery at 18:18 UTC but was still investigating the root cause; no specific fix was stated.

Cursor reported a major service degradation in which Automations, Cloud Agents, and Review Agents failed to start. The incident began at 2026-09-02 14:50 UTC and was resolved at 2026-09-02 16:10 UTC, lasting 1 h 20 min; no cause or fix was stated.

Gorgias reported a major degradation delaying Instagram direct message ingestion from 2026-09-02 14:49 UTC to 2026-09-02 19:11 UTC, lasting 4 h 22 min. Messages were not lost but arrived later than usual. A fix was implemented at 18:03 UTC, and the incident was resolved.

Buildkite reported a major Hosted Agents job-dispatch degradation affecting Linux (AMD64), Linux (ARM64), and MacOS from 2026-09-02 14:45 UTC to 2026-09-02 15:51 UTC, lasting 1 h 5 min. Self-hosted agents were unaffected; Buildkite worked through the delayed-job backlog and confirmed dispatch delays were resolved.

Xero’s UK Making Tax Digital (MTD) and VAT services had a critical outage starting 2026-09-02 14:05 UTC, lasting 3 h 48 min. HMRC investigated; the incident was marked resolved at 2026-09-02 17:53 UTC, although the final update also said services were “not operating as normal.”

Chargebee reported major API degradation affecting some US-region users, with intermittent failures between 12:54 UTC and 12:57 UTC. The incident record started at 2026-09-02 13:28 UTC and was resolved at 2026-09-02 13:35 UTC, lasting 7 min; services had recovered by the initial update and were monitored for stability.

GitLab reported a major GitLab.com CI/CD degradation affecting instance runners from 2026-09-02 11:54 UTC to 2026-09-02 12:43 UTC (49 min), delaying job pickup, pipelines and deployments. A manually initiated pg_repack maintenance operation caused CI database contention affecting runner communication; GitLab terminated the operation, runner metrics recovered, and the incident was resolved.

Greenhouse reported a major partial disruption affecting job creation and causing multiple confirmation emails to be sent to candidates on Greenhouse Recruiting - Silo 201 and Silo 201. It started at 2026-09-02 08:06 UTC and lasted 26 min, resolving at 2026-09-02 08:32 UTC after a fix was implemented; all services returned to normal.

Optimizely reported a major SSL trust outage affecting DXP environments in Australia East and East US, preventing affected customers from logging into the CMS management interface or creating/editing content. It lasted 7 h 18 min, from 2026-09-02 08:05 UTC to 2026-09-02 15:23 UTC; Optimizely implemented a fix and confirmed normal service.

Fly.io reported a major API background job queue failure starting 2026-09-02 07:21 UTC: creating apps, assigning IP addresses, and creating/renewing certificates could fail. A fix was implemented by 07:31 UTC; the incident was resolved at 07:44 UTC, lasting 22 min.

Users could not access billing pages in their Buffer accounts during a major-impact incident starting 2026-09-02 06:30 UTC. Buffer deployed a fix at 07:28 UTC, restoring page loading, and resolved the incident at 2026-09-02 07:57 UTC; the incident lasted 1 h 27 min.

Optimizely reported a critical CMP and DAM outage affecting all users, starting 2026-09-02 05:00 UTC and resolved at 2026-09-02 07:49 UTC, lasting 2 h 48 min. A fix was implemented by 06:44 UTC, when both products were reported accessible; sustained stability was confirmed at closure. The root cause was identified but not disclosed.

QuickBooks Online reported a major incident affecting some users’ logins and causing errors within application features across APAC, Canada, EMEA and the United States. The recorded start was 2026-09-02 04:16 UTC; it lasted 3 min and was resolved at 2026-09-02 04:19 UTC after a fix, with recovery confirmed across all regions.

Elastic Cloud reported a major incident with elevated error rates for Anthropic Claude Haiku 4.5, OpenAI GPT-5.6, and Google Gemini 2.5 Flash on Elastic Inference Service in EU regions, citing 21:50–23:42 UTC. It was resolved; recorded start and resolution were both 2026-09-02 00:04 UTC (computed duration: 0 min).

Vercel reported a major incident causing some deployments using IAD1 Function regions or Routing Middleware to fail. It started at 2026-09-01 19:59 UTC and lasted 59 min, resolving at 2026-09-01 20:58 UTC after a fix was deployed and Builds recovered.

v0 reported a major incident causing some deployments using IAD1 Function regions or Routing Middleware to fail. It lasted 59 min, from 2026-09-01 19:59 UTC to 2026-09-01 20:58 UTC; a fix was deployed and Builds recovered.

Brex’s Trust Portal (trust-portal.brex.com) was unavailable in a major-impact incident starting 2026-09-01 17:48 UTC. Brex reported the site restored at 09-01 19:20 UTC and marked the incident resolved at 2026-09-02 01:56 UTC; the recorded incident duration was 8 h 7 min.

Anaconda reported a critical network interruption affecting the anaconda.org front-end and conda.anaconda.org repository services, with elevated 503 errors on anaconda.org. It started at 2026-09-01 17:46 UTC and was resolved at 2026-09-01 19:05 UTC, lasting 1 h 18 min. A fix was rolled out at 18:53 UTC; no cause was stated.

Gorgias reported a major Voice degradation starting 2026-09-01 17:05 UTC due to a Twilio incident, causing failed or delayed calls and inaccurate call test results for some customers. SMS, email, chat and helpdesk were unaffected. Resolved at 2026-09-01 18:00 UTC after 54 min, Voice functionality and call test results returned to normal.

Matillion reported a major Hub login outage starting 2026-09-01 17:03 UTC, resolved at 2026-09-01 17:55 UTC after 52 min; the support portal was also affected. New tokens could not be issued after hitting the global token limit; Matillion attributed widespread permission-check failures to an out-of-memory event triggered by a Permission Service deployment.

Apollo reported a major outage from 2026-09-01 15:41 UTC to 2026-09-01 20:18 UTC (4 h 37 min), attributed to GCP issues in part of us-central1-b. New schemas could not be published; AWS Uplink remained available, allowing default multi-cloud GraphOS Routers and Gateways to start and fetch configuration. All Apollo products recovered.

Sourcegraph reported a major incident affecting account sign-in and authentication, starting 2026-09-01 15:41 UTC and resolved 2026-09-02 09:39 UTC, lasting 17 h 58 min. Sourcegraph attributed the errors to a Google Cloud networking incident; sign-in had recovered by 2026-09-01 16:48 UTC, with monitoring continuing afterward.

CockroachDB Basic/Standard experienced major latency, delays, and connection issues affecting us-central-1, asia-east-1 (later listed as us-central1, asia-east1), europe-west4-b, and us-central1-b from 2026-09-01 15:36 UTC to 2026-09-01 19:45 UTC (4 h 8 min). CockroachDB attributed the disruption to Google Cloud networking outages; network-level mitigation restored infrastructure, and all services were confirmed restored.

A critical outage prevented GCP-hosted organizations in us-central1 from reliably accessing Sigma. The recorded incident ran from 2026-09-01 15:34 UTC to 2026-09-01 16:55 UTC (1 h 20 min), though Sigma reported an approximate start of 2:46 PM UTC. The postmortem attributed the disruption to a Google Cloud networking incident; services returned to normal.

Dialpad reported a major partial service disruption affecting Telephony Infrastructure: Inbound Calls, starting 2026-09-01 15:25 UTC and resolved at 2026-09-01 20:02 UTC, lasting 4 h 36 min. Dialpad stated that a fix was implemented but did not specify the cause or fix details.

Gorgias reported a major cloud service provider network incident potentially affecting a subset of customers, with Live Chat, REST API and Google Cloud Networking listed as affected components. It started at 2026-09-01 15:19 UTC and was resolved at 2026-09-01 18:59 UTC, lasting 3 h 39 min.

Fullstory reported a critical NA1 outage affecting data capture (web and native mobile), the Fullstory web application, API and other services. It ran from 2026-09-01 15:19 UTC to 2026-09-01 20:36 UTC (5 h 17 min). Fullstory linked the issue to Google Cloud Platform; services were restored, with backlogged data still processing.

A major outage impacted Elastic Cloud Hosted and Serverless projects in GCP us-central1 and the Docker Registry service. It started at 2026-09-01 15:18 UTC and was resolved at 2026-09-01 20:05 UTC, lasting 4 h 47 min; Elastic attributed the disruption to an upstream CSP issue.

Drata reported a critical outage affecting Trust Center Pages, Trust Center Admin Page, AIQA, and Compliance Monitoring (AI Summary), caused by upstream provider network service degradation. It started at 2026-09-01 15:14 UTC and was resolved at 2026-09-01 19:29 UTC, lasting 4 h 14 min; all impacted services were operational at resolution.

Pinecone reported a major incident causing 5xx errors on some control plane operations, affecting Console, Index management, Assistant management, and APIs at api.pinecone.io. It started at 2026-09-01 15:04 UTC and was resolved at 2026-09-01 19:40 UTC, lasting 4 h 36 min. Pinecone said it was likely related to ongoing GCP issues.

RevenueCat reported a major Apple Receipt Server degradation from 2026-09-01 13:17 UTC to 2026-09-01 13:43 UTC (26 min), with intermittent HTTP 522 responses from its receipts endpoint. Mitigations meant no purchases were being blocked; Apple reported resolution, error rates normalized, and RevenueCat said it was ensuring affected purchases were reprocessed.