Notable SaaS vendor changes, page 16

Notion reported a major outage affecting access to notion.com and Notion Sites, with possible issues loading or using Notion clients. The incident started at 2026-08-06 00:29 UTC and was resolved at 2026-08-06 00:37 UTC (8 min tracked duration; the update reported 14 minutes of inaccessibility). Notion reverted the change that inadvertently triggered the outage.

Elastic Cloud reported a major Elasticsearch 9.5.0 defect: boolean queries combining must, filter, or should with must_not on unindexed fields could return false positives and inflated counts. It lasted from 2026-08-05 22:59 UTC to 2026-08-12 08:59 UTC (6 days 10 h); Elasticsearch 9.5.1 fixed it, and customers running 9.5.0 were advised to upgrade.

HashiCorp reported stuck Infragraph snapshots and degraded performance, with major impact, starting at 2026-08-05 20:09 UTC. The incident lasted 1 h 58 min and was resolved at 2026-08-05 22:08 UTC; engineering said systems should be operating normally, but provided no cause or specific fix.

Terraform Cloud (HCP) reported stuck Infragraph snapshots and degraded performance, with major impact, starting 2026-08-05 20:09 UTC. The incident lasted 1 h 58 min and was resolved at 2026-08-05 22:08 UTC; engineering said systems should be operating normally, but provided no cause or specific fix.

Cohere reported a major partial outage of its documentation site: some pages failed to load correctly due to an issue with upstream documentation provider Fern. The incident started at 2026-08-05 19:36 UTC and was resolved at 2026-08-05 20:35 UTC, lasting 58 min; Cohere confirmed service was back to normal.

Descript reported a critical Rooms outage starting at 2026-08-05 19:29 UTC, lasting 28 min and resolved at 2026-08-05 19:57 UTC. Functionality was fully restored at 19:32 UTC, followed by monitoring; no cause was stated.

Terraform Cloud (HCP) reported a critical outage of the Infragraph Service and UI starting 2026-08-05 15:11 UTC. Engineering implemented a fix and confirmed normal operation at 2026-08-05 15:39 UTC; the reported duration was 27 min.

HashiCorp reported a critical outage of the Infragraph Service and UI starting 2026-08-05 15:11 UTC, with degraded performance reported. Engineering implemented a fix at 2026-08-05 15:28 UTC and marked the incident resolved at 2026-08-05 15:39 UTC; the computed duration was 27 min.

Dremio Cloud US experienced a major outage affecting application loading and login; Dremio Cloud Europe operated normally. The incident started at 2026-08-05 14:47 UTC and was resolved at 2026-08-05 18:49 UTC, lasting 4 h 2 min; Dremio reported service recovery and healthy services while monitoring.

ActiveCampaign reported a critical outage affecting segmented campaign sending in the EU Data Center, starting 2026-08-05 10:59 UTC and resolved 2026-08-06 11:40 UTC, lasting 24 h 41 min. A fix was implemented at 08-05 13:12 UTC; no cause was stated.

Apollo GraphQL’s Embedded Explorer and Embedded Sandbox were down in a major-impact incident from 2026-08-05 09:43 UTC to 2026-08-05 10:08 UTC, lasting 25 min. Apollo identified the issue, deployed a fix, and confirmed resolution; no cause was stated.

Coralogix reported a major incident causing log query failures in Explore → Logs (v2 and Legacy), including errors accessing archived log data. It started at 2026-08-05 08:46 UTC and was resolved at 2026-08-05 09:01 UTC, lasting 15 min. Although initially reporting EU1 and US1 impact, Coralogix confirmed only EU1 was affected; all services were restored.

MongoDB Atlas reported a major incident affecting some customers’ deployments due to GCP capacity constraints across North America, Europe, and Asia Pacific/Australia, including M50 and S70L instance types. It started at 2026-08-05 05:17 UTC and was resolved at 2026-08-05 12:14 UTC, lasting 6 h 56 min; MongoDB said symptoms had largely subsided and Atlas was fully operational.

New Claude-to-Notion MCP authorizations failed with “Invalid authorization request,” blocking new connections. Notion rated the incident major and reported failures starting approximately 01:25 UTC on August 5, 2026; the incident was logged at 2026-08-05 03:34 UTC and resolved at 2026-08-05 04:33 UTC (computed duration: 58 min).

Greenhouse reported a major partial disruption in Greenhouse Recruiting - Silo 1, affecting user searches, candidate unhire actions, emails, availability requests, referrals, offer letters, scorecards, candidate notes and interview-plan editing. It started 2026-08-05 02:18 UTC and resolved 2026-08-05 13:29 UTC (11 h 10 min); a fix was implemented at 08-05 03:56 UTC.

Convex reported a major incident causing actions that call createFunctionHandle to fail on business deployments; Free, Starter and Professional users were unaffected. It started at 2026-08-05 01:42 UTC and was resolved at 2026-08-05 01:57 UTC, lasting 14 min. Convex reported implementing and deploying a fix before confirming resolution.

HCP Terraform had a major UI outage causing blank white screens from 2026-08-05 00:05 UTC to 2026-08-05 01:59 UTC, lasting 1 h 54 min. API calls, CI integrations, and CLI usage were not expected to be affected. Engineering implemented a fix at 00:44 UTC and confirmed resolution at 01:59 UTC.

HCP Terraform’s UI showed a blank white screen in a major outage from 2026-08-05 00:05 UTC to 2026-08-05 01:59 UTC, lasting 1 h 54 min. API calls, CI integrations, and CLI usage were not expected to be affected. Engineering implemented a fix at 00:44 UTC and confirmed resolution at 01:59 UTC.

CockroachDB reported a major outage preventing some Basic and Standard clusters in eu-west1 from connecting or executing queries. It ran from 2026-08-04 23:46 UTC to 2026-08-05 19:50 UTC (20 h 3 min). Immediate impact was mitigated by 2026-08-05 03:55 UTC; resolution followed monitoring confirming stability, with no root cause stated.

Anthropic reported a major incident affecting claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork: some users had login issues, including OAuth, and model request errors. It started at 2026-08-04 20:48 UTC and was resolved at 2026-08-04 21:59 UTC, lasting 1 h 11 min; no specific cause or fix was stated.

Claude reported a major incident affecting claude.ai, Claude API (api.anthropic.com), Claude Code and Claude Cowork: some users had login issues, including via oAuth, and model request errors. It started at 2026-08-04 20:48 UTC and was resolved at 2026-08-04 21:59 UTC, lasting 1 h 11 min; success rates had recovered by 21:04 UTC.

Box reported major impact to uploads and downloads across its API and web application, with errors and slowness affecting customers. The incident ran from 2026-08-04 20:23 UTC to 2026-08-04 20:44 UTC (21 min). Box attributed upload failures and delays to frontend load-balancer routing and restored service by rerouting upload traffic and provisioning additional capacity.

Some Grafana instances in GCP US Central (prod-us-central-0, prod-us-central-3, prod-us-central-4) were unavailable during a critical incident from 2026-08-04 19:11 UTC to 2026-08-05 01:32 UTC (6 h 20 min). Grafana reported cloud-provider compute capacity exhaustion, moved affected workloads to alternative capacity, and rolled out a change to restore service.

Fly.io reported a major outage in GRU: new MPG clusters could fail to create due to insufficient capacity. It started at 2026-08-04 19:03 UTC and was resolved at 2026-08-05 03:22 UTC, lasting 8 h 19 min; Fly.io tweaked hosts to allow more machine allocation.

Discord reported a critical API outage affecting direct messages, interactions, shop, and SDK integrations from 2026-08-04 18:49 UTC to 2026-08-04 20:17 UTC, lasting 1 h 27 min. A fix was in place by 19:53 UTC, with nearly all users recovered by 19:59 UTC; the incident was resolved at 20:17 UTC.

OneSignal reported a major incident causing about 3% of Apple push notifications to fail, from 2026-08-04 17:20 UTC to 2026-08-04 19:24 UTC (2 h 4 min). The vendor attributed it to a slight change in how Apple's push notification connections work, applied a mitigation, and confirmed normal operation.

Brex reported a critical incident affecting API requests, starting 2026-08-04 16:01 UTC and resolved at 2026-08-04 20:34 UTC, lasting 4 h 32 min. Brex stated a fix was in place at 08-04 17:50 UTC; no cause was disclosed.

Wasabi reported a major S3 outage in Frankfurt (eu-central-2) from 07:36 UTC to 08:03 UTC on 4 August 2026, caused by a frontend software update; rollback restored service. The record lists start 2026-08-04 14:20 UTC and resolution 2026-08-04 07:30 UTC, conflicting with the update; computed duration is unknown.

Docker reported a major outage: Git pushes to Bitbucket repositories did not trigger automated builds; GitHub repositories were unaffected. The incident ran from 2026-08-04 14:12 UTC to 2026-08-04 16:09 UTC, lasting 1 h 57 min. A fix was deployed at 15:42 UTC, and Bitbucket build triggers were confirmed restored at resolution.

JFrog reported a major incident affecting Artifactory in Azure EU West (Netherlands), with some customers potentially experiencing a brief service interruption. It started and was resolved at 2026-08-04 13:24 UTC, with a computed duration of 0 min; JFrog validated service stability and continued monitoring.

Snyk reported a major incident affecting all Snyk Open Source .NET project workflows, including import, retest, and pull request checks, across AU, EU, GOV, US-01 and US-02 components. It started at 2026-08-04 13:07 UTC and was resolved at 2026-08-04 14:29 UTC, lasting 1 h 21 min; engineers deployed a fix and confirmed recovery.

Socket reported major elevated latency affecting Socket for GitHub, Package Pages, Socket API, Socket Dashboard and Socket Website, starting 2026-08-04 07:42 UTC. A fix was implemented by 2026-08-04 08:33 UTC, and the incident was resolved at 2026-08-04 08:45 UTC after 1 h 3 min.

Lovable reported a critical outage affecting TanStack application previews and published apps; previews could show an internal server error. The incident started at 2026-08-04 00:27 UTC and was resolved at 2026-08-04 01:52 UTC, lasting 1 h 25 min; Lovable reported all systems operating normally.

VoyageAI APIs experienced elevated 5xx errors affecting VoyageAI Embedding and Reranking Models, with major impact, from 2026-08-03 18:47 UTC to 2026-08-03 21:18 UTC (2 h 31 min). A fix was applied at 19:02 UTC, and normal operation was confirmed at resolution.

PandaDoc’s critical live chat outage affected Support chat and US & Global - Support chat, starting 2026-08-03 15:18 UTC and lasting 2 h 36 min. PandaDoc attributed it to an Intercom provider outage and warned support replies could be delayed; live chat was confirmed working and the incident resolved at 2026-08-03 17:55 UTC.

Snowflake’s critical outage affected AWS - US East (Ohio) - Virtual Warehouses, blocking warehouse starts, resumes or management. Reported onset: 13:35 UTC August 03, 2026; tracked start: 2026-08-03 14:11 UTC; resolved: 2026-08-03 20:58 UTC; recorded duration: 6 h 47 min. Snowflake applied an additional fix after a Cloud Services component became unhealthy again.

Netlify reported major elevated build errors affecting the Build Pipeline, with errors beginning at 8:58 UTC and the incident recorded from 2026-08-03 09:31 UTC. A fix was implemented, and the incident was resolved at 2026-08-03 09:55 UTC; the recorded duration was 24 min.

Buildkite reported a major Job Queue degradation starting 2026-08-03 05:11 UTC, resolved at 2026-08-03 05:36 UTC after 25 min. Scheduled builds were delayed, alongside job dispatch, pipeline uploads and incoming webhooks. A fix was fully deployed at 05:29 UTC; Buildkite said affected events would be automatically retried.

Optimizely Analytics (Netspring) experienced a major outage affecting Netspring customers only, starting 2026-08-03 02:30 UTC and resolved at 2026-08-03 16:04 UTC, lasting 13 h 33 min. Optimizely reported identifying the underlying cause without disclosing details and confirmed service was restored to normal operation.

Plaid reported a major disruption affecting new connections to Bank of America, starting 2026-08-02 14:51 UTC and resolved at 2026-08-03 12:53 UTC, lasting 22 h 1 min. Plaid notified Bank of America; no cause or specific fix was stated.

Wix reported a critical outage affecting services including Wix Editor, Sign-In, Checkout, Wix Payments, APIs, Bookings and Dashboard, impacting site management and customer transactions. It started at 2026-08-01 11:27 UTC and was resolved at 2026-08-01 14:44 UTC; no cause or fix was stated.

incident.io reported a major degradation affecting some AI features, with increased error rates starting at 2026-08-01 00:08 UTC. The incident lasted 2 h 37 min and was resolved at 2026-08-01 02:45 UTC, when the vendor reported full recovery.

Temporal Cloud’s Usage API and UI had a critical outage affecting Billing and Global Services - Billing, starting 2026-07-31 22:46 UTC and lasting 2 h 13 min. A fix was implemented at 2026-08-01 00:30 UTC, and the incident was resolved at 2026-08-01 01:00 UTC; no cause was stated.

Canva reported a major incident affecting subscriptions, billing and payments from 2026-07-31 18:55 UTC to 2026-07-31 19:58 UTC, lasting 1 h 3 min. Some users had payment issues; Canva implemented a fix at 19:24 UTC and confirmed resolution at 19:58 UTC.

Optimizely reported a major CMP outage: the Home Page, Plan module, and Requests module failed to load for some customers. It started at 2026-07-31 17:53 UTC and was resolved at 2026-07-31 20:10 UTC, lasting 2 h 17 min. A fix was implemented; no cause was stated.

WorkOS reported a major AuthKit SSO incident causing elevated errors and sign-on redirect loops for a small subset of customers; standalone SSO was unaffected. It started at 2026-07-31 16:12 UTC and resolved at 2026-07-31 18:06 UTC, lasting 1 h 53 min. A fix was deployed at 17:10 UTC.

Some Lovable project builds failed in a major-impact incident starting 2026-07-31 12:24 UTC, lasting 34 min and resolved at 2026-07-31 12:59 UTC. Lovable said its 3rd party provider fixed the issue and that it monitored recovery.

Box reported a major-impact outage affecting Box Shuttle Standalone: users could have trouble accessing, configuring or creating Shuttle job details. It started at 2026-07-31 11:09 UTC and was resolved at 2026-07-31 11:30 UTC, with a computed duration of 20 min. Box reported remediation steps and restored full functionality, without specifying a cause.

ElevenLabs reported a major SIP outage affecting calls to some numbers imported into the Agent Platform, starting 2026-07-31 00:45 UTC and lasting 18 h 15 min. The vendor cited infrastructure resource constraints as a likely contributor; remaining failures affected the US region. Mitigation returned failure rates to baseline, and the incident was resolved at 2026-07-31 19:00 UTC.