Major incidents SaaS vendor changes, page 18

Dekaf materializations stalled on public AWS us-east-1 and GCP us-central1 data planes in a major incident from 2026-07-22 11:19 UTC to 2026-07-22 18:06 UTC, lasting 6 h 46 min. All other tasks and data planes were unaffected. Estuary deployed a fix and resolved the incident; no customer action was required.

Snyk reported major Open Source and Container degradation on SNYK-US-01 (app.snyk.io), starting 2026-07-22 09:41 UTC and resolved after 21 h 28 min. Customers may have experienced import errors and Project details issues from 09:08 to 09:30 UTC; corrective actions restored operations. No further impact was observed during monitoring; customers should re-import failed imports.

Prefect Cloud (US) experienced a critical incident affecting API requests, UI, and orchestration, with delayed flow run scheduling, automations, and event processing. It started at 2026-07-22 09:31 UTC and was resolved at 2026-07-22 09:58 UTC, lasting 26 min. Prefect reported resolving an underlying database issue; all services, including background processing, returned to normal.

Wix reported a major outage affecting My Sites, Sign-In, Site Dashboard, and site editing through Wix Editor and Wix Studio, disrupting access and editing. It started at 2026-07-22 08:36 UTC and was resolved at 2026-07-22 08:48 UTC; the computed duration was 11 min. Wix implemented a fix but did not state the cause.

Docker Hub experienced intermittent timeouts on hub.docker.com, with major impact, from 2026-07-22 05:34 UTC to 2026-07-22 08:00 UTC, lasting 2 h 26 min. Docker implemented a fix, then investigated incomplete recovery before confirming normal operations.

WorkOS reported a major incident involving SSO failures for AuthKit customers, starting at 2026-07-21 21:21 UTC and resolved at 2026-07-21 21:54 UTC, lasting 32 min. WorkOS said it identified and fixed the issue and AuthKit SSO fully recovered; no cause was stated.

Klaviyo reported a major incident affecting SMS subscriptions, delaying SMS sends, and causing errors creating or scheduling SMS and WhatsApp campaigns. It started at 2026-07-21 20:36 UTC and was resolved at 2026-07-21 21:09 UTC, lasting 32 min. Klaviyo implemented a fix but did not disclose the cause.

Pinecone reported a critical incident causing 5xx errors for some Serverless namespaces in Azure eastus2. It started at 2026-07-21 19:10 UTC and was resolved at 2026-07-21 19:40 UTC, with a reported duration of 29 min; Pinecone identified the issue and implemented a fix but did not specify the cause.

SquadCast recordings intermittently failed in a major-impact incident starting 2026-07-21 19:01 UTC and lasting 39 min, resolved at 2026-07-21 19:40 UTC. Descript reported functionality fully restored at 2026-07-21 19:14 UTC but did not state a cause or specific fix.

OpenAI reported a major outage affecting ChatGPT and API Platform logins and account creation, plus API Platform organization management and dashboard access. It started at 2026-07-21 18:36 UTC and was resolved at 2026-07-21 19:00 UTC, lasting 23 min; the final update confirmed full recovery of affected API Platform functionality.

A major outage affected logins and account creation for ChatGPT and the API Platform, plus API Platform organization management and dashboard access; users already signed in may still have been able to use ChatGPT. It started at 2026-07-21 18:36 UTC and was resolved at 2026-07-21 19:00 UTC (23 min), with all affected API Platform functionality fully recovered.

API image generation requests experienced elevated error rates in a major-impact incident from 2026-07-21 18:10 UTC to 2026-07-21 22:19 UTC, lasting 4 h 9 min. OpenAI applied mitigations and confirmed full recovery, with error rates back to normal.

API image generation requests experienced elevated error rates in a major-impact incident from 2026-07-21 18:10 UTC to 2026-07-21 22:19 UTC, lasting 4 h 9 min. OpenAI applied mitigations and confirmed full recovery, with error rates back to normal.

Attio reported a major workflow-triggering incident starting 2026-07-21 17:35 UTC, resolved after 47 h 53 min. Some workflows failed between 16:33 and 18:01 UTC on July 21, 2026, because an infrastructure configuration change caused encryption errors. Attio reverted the change, corrected the configuration, and notified impacted customers.

Elastic Cloud reported a major Instance Capacity metering issue affecting a small number of Hosted Deployments customers, starting 2026-07-21 14:24 UTC and lasting 32 h 50 min. Resolved at 2026-07-22 23:14 UTC, all metering errors were corrected, and Elastic said it would issue refunds to affected customers.

Elastic Cloud reported major AutoOps degradation: a subset of Elastic Agents could not report metrics across all regions during Jul 18 02:40 UTC–Jul 20 14:00 UTC, leaving monitoring gaps. The issue was corrected and data resumed normally. Posted retroactively, the resolved incident record started 2026-07-21 13:45 UTC, with a computed duration of 35 min.

Matillion reported a critical Hub outage preventing projects from loading within Maia Foundation, starting 2026-07-21 11:39 UTC. The incident lasted 1 h 17 min and was resolved at 2026-07-21 12:56 UTC after a fix was implemented; no cause was stated.

ChatGPT image generation was unavailable with elevated errors in an incident rated major, starting 2026-07-21 10:36 UTC and lasting 6 days 8 h. OpenAI reported applying a mitigation and monitoring recovery; the final resolution at 2026-07-27 19:04 UTC stated that all impacted services had fully recovered.

ChatGPT image generation was unavailable with elevated errors in a major-impact incident starting 2026-07-21 10:36 UTC and resolved at 2026-07-27 19:04 UTC (recorded duration: 6 days 8 h). OpenAI applied a mitigation and reported recovery, followed by renewed errors before final recovery.

Elastic Cloud reported a major incident causing delayed or failed email MFA code delivery for Gmail users, starting 2026-07-21 09:36 UTC and resolved 2026-07-23 16:37 UTC (2 days 7 h). Delays were mitigated on 07-21 23:10 UTC; customers could use registered TOTP/WebAuthn factors or “Log in with Google” as workarounds.

dbt Labs reported a critical outage from 2026-07-21 07:45 UTC to 2026-07-21 12:52 UTC (5 h 7 min). Scheduled, CI, and API-triggered GitHub job runs failed at “Clone git repository” across all regions (multi-tenant and single-tenant) because of upstream GitHub SSH deploy key authentication failures. GitHub resolved the issue; failed runs can be retried.

Heap reported a major Heap Connect incident affecting some US Connect customers with delays in data lake ingestion. It started at 2026-07-21 06:34 UTC and was resolved at 2026-07-21 09:06 UTC, lasting 2 h 31 min; Heap identified the issue and implemented a fix, but did not state a cause.

Sentry reported a critical EU-region incident causing high ingestion latency for errors, attachments, and feedback from 2026-07-20 11:11 UTC to 2026-07-20 16:57 UTC, lasting 5 h 46 min. Sentry implemented an attachment ingestion fix and cleared attachment and error backlogs; latency returned to normal across all event types.

Fly.io reported a critical outage with high numbers of 5XX errors on the Machines API and dashboard, starting 2026-07-20 07:10 UTC and resolved at 2026-07-20 17:09 UTC after 9 h 59 min. A fix was implemented; some Managed Postgres v1 clusters were also degraded, while existing machines and Managed Postgres v2 were unaffected.

ChatGPT reported a major incident affecting conversation loading and continuation, Voice, and Work Mode, starting 2026-07-19 14:49 UTC and lasting 2 h 15 min. Mitigation was applied, and all impacted services were fully recovered by 2026-07-19 17:05 UTC.

OpenAI reported a major ChatGPT incident affecting conversation loading and continuation, Voice, and Work Mode. It started at 2026-07-19 14:49 UTC and resolved at 2026-07-19 17:05 UTC, lasting 2 h 15 min. OpenAI applied a mitigation and confirmed full recovery; the specific cause was not disclosed.

Hootsuite reported a major Meta-related degradation affecting Facebook and Instagram message delivery to Inbox and data completeness in Analytics. It started at 2026-07-19 09:19 UTC and was resolved at 2026-07-19 09:50 UTC, lasting 30 min; customers with lingering missing messages or incomplete data were directed to support.

Buffer reported a major outage affecting video publishing to Instagram and Threads from 2026-07-19 09:06 UTC to 2026-07-19 10:01 UTC, lasting 55 min, due to a problem with Meta’s API. Image and text-only posts were unaffected. Meta implemented a fix; publishing was restored, and customers can retry failed posts from the queue.

Major Git clone failures (403 errors) affected EMEA scheduled and triggered jobs using the GitHub App integration; IDE and CLI were unaffected. Failures began 2026-07-18 23:00 UTC; the incident was tracked from 2026-07-19 07:33 UTC until resolution at 2026-07-19 13:55 UTC (6 h 22 min). Errors normalized without a deployed fix; no definitive root cause was identified.

Brex reported major-impact scheduled maintenance starting 2026-07-19 06:32 UTC, lasting 1 h 13 min and resolving at 2026-07-19 07:45 UTC. The notice warned of brief dashboard interruptions and unavailable new signups, banking transactions, and card creation; Brex confirmed maintenance was completed.

Make reported a major US2 incident preventing new connection creation and connection re-authorization with TypeError (SC424), affecting Apps - us2.make.com and us2.make.com. It started at 2026-07-19 05:58 UTC, lasted 26 h 51 min, and was marked resolved at 2026-07-20 08:49 UTC; no cause or fix was specified.

Webflow reported a major degradation affecting hosted pages with CMS content, with intermittent 5xx errors during several brief windows. The incident started at 2026-07-19 05:14 UTC and resolved at 2026-07-19 14:00 UTC (8 h 45 min); Webflow confirmed its mitigation restored expected site performance.

Render reported a critical incident: GitHub git clone 403 errors blocked deployments for services in Frankfurt; running services were unaffected. It lasted 5 h 36 min, from 2026-07-19 03:44 UTC to 2026-07-19 09:20 UTC, when it was resolved; no specific cause or fix was stated.

Baseten’s GLM 5.2 Model API experienced a major partial outage due to provider network failure, starting 2026-07-18 22:40 UTC and lasting 32 min until resolution at 2026-07-18 23:12 UTC. Dedicated inference was not affected; Baseten reported implementing a fix at 07-18 23:05 UTC.

Brex’s major-impact scheduled maintenance started July 18 at 07:01 UTC and was completed at 09:32 UTC, lasting 2 h 31 min. Customers were warned of brief interruptions accessing accounts via the dashboard or mobile app, and temporary unavailability of adding Brex cards to digital wallets such as Apple Pay or Google Pay.

Optimizely reported a major partial outage of Campaign’s graphical interfaces for managing campaigns and content, including the EMEA Region interface, affecting only some users. It started at 2026-07-17 16:59 UTC and was marked resolved at 2026-07-17 17:17 UTC; the computed duration was 17 min. The vendor identified the affected component and resolved the problem.

Brex reported a critical Spend Management incident affecting access to new spend limit and spend limit increase requests on the dashboard’s Tasks page. It started at 2026-07-17 14:51 UTC and was resolved at 2026-07-17 15:32 UTC, lasting 40 min. Brex said it identified the root cause and resolved the issue but provided no details.

Make reported a major login outage on make.com starting 2026-07-17 14:22 UTC, resolved at 2026-07-17 15:09 UTC after 47 min. Scenario executions and webhooks were unaffected. Make identified the issue and applied a fix at 07-17 14:28 UTC, but did not state the cause.

Make reported a major outage affecting Etsy default connections, which returned an “invalid_target” error. It started at 2026-07-17 13:34 UTC and was resolved at 2026-07-22 12:07 UTC, lasting 4 days 22 h. Make attributed the issue to Etsy’s API and reported resolution after a temporary outage on Etsy’s side.

Bunny.net reported a major Dashboard login outage from 2026-07-17 10:57 UTC to 2026-07-17 11:11 UTC, lasting 14 min. Users could not log in to dash.bunny.net, but services themselves were not disrupted. The issue was resolved after the Frontend team deployed a fix.

Make reported a major login outage across multiple zones starting 2026-07-17 10:20 UTC, resolved at 2026-07-17 12:21 UTC, with a reported duration of 2 h 0 min. Scenario executions and webhooks were unaffected; Make applied a fix at 10:27 UTC and monitored results before resolving the incident.

Discord reported a major API Errors incident from 2026-07-17 00:49 UTC to 2026-07-17 03:48 UTC, lasting 2 h 59 min. Discord identified an upstream latency issue, worked to mitigate its impacts, and marked the incident resolved.

Socket’s GitHub App reports failed in a critical-impact incident starting 2026-07-16 23:45 UTC, lasting 15 min and resolved at 2026-07-17 00:00 UTC. Socket attributed the failures to a high error rate on GitHub’s API and reported that GitHub restored service.

Render reported a critical GitHub outage affecting new service creation and service deploys in Frankfurt, Ohio, Oregon, Singapore, and Virginia. It started at 2026-07-16 22:42 UTC and was resolved at 2026-07-17 00:14 UTC, lasting 1 h 32 min. GitHub reported degraded REST API availability and later said it had mitigated its incident.

Box reported a major-impact outage affecting Box Edit and Microsoft Office Online, starting 2026-07-16 21:30 UTC and resolved at 2026-07-16 21:48 UTC (18 min). Box reported full functionality restored after remediation; its postmortem said some customers may have been unable to open or edit files through Open With integrations in the Box web application.