Tuesday 1 Sept · 20:05 UTC · updated 595 min ago

Drama index 50/10012 degraded · 4 downMood: Unsettled

Fireworks AI: drizzle on Serverless models

Immediate amendment: Fireworks AI has opened a Serverless-model degradation file; routine automation is asked to await further notice.

Apple’s subscription-purchase wobble has cleared; the small print survives.Apple’s subscription-purchase wobble has cleared; the small print survives.Apple’s subscription-purchase wobble has cleared; the small print survives.Apple’s subscription-purchase wobble has cleared; the small print survives.Apple’s subscription-purchase wobble has cleared; the small print survives.Hetzner tests UPS and emergency power in FSN1-DC20; suspense is operational.

Should you deploy right now?

Ship itAbsolutely not

Use caution. Vercel is down; Cloudflare and Supabase are degraded.

Deployment inputs: 1 down, 2 degraded · no unconfirmed warnings.

Excuse of the hour — Cloudflare’s “Requests from wrangler or other OAuth clients may experience elevated errors” makes OAuth a respectable suspect.

VercelBlackoutIncreased deployment failuresDetails →

RedditThunderElevated site errorsDetails →

SourcegraphBlackoutSign-in and authentication errors+2 moreDetails →

Elastic CloudThunderIssue impacting services running in GCP us-central1+1 moreDetails →

Fireworks AIHazyService Degradation for one of our Serverless models+1 moreDetails →

ZoomDrizzleService Degradation Affecting Zoom for Google Workspace add-onDetails →

Atlassian JiraSlowDegraded performance of JIRADetails →

SentryDrizzleSpan ingestion is degraded in USDetails →

SalesforceDrizzleSalesforce service incidentDetails →

TwilioDelayedSMS Delivery Delays from Twilio to Orange Democratic Republic of the Congo+7 moreDetails →

CloudflareDelayedHTTP/3 issue affecting R2 custom domains+1 moreDetails →

CoinbaseSlowPaused Sends/Receives - Core DAO NetworkDetails →

KrakenHazyCronos (CRO) Funding Disruptions+1 moreDetails →

SnowflakeDelayedINC20000150Details →

falDrizzleHistorical Request lookup and Assets generated between Aug 1 - 26 may be degradedDetails →

+ 1 more affected →

Supabasedegraded

All clear · 129 services

Regional

AWS · 9 regions

us-east-1 ap-northeast-1 ap-south-1 ap-southeast-1 eu-central-1 eu-west-1 me-central-1 sa-east-1 us-west-2

Quiet · 37 community services with no reports

Forecast

Fireworks AI has entered a service degradation for one of its Serverless models, the latest amendment to a crowded operational register. Vercel remains down with increased deployment failures, while Elastic Cloud and Reddit are also listed down. Cloudflare’s HTTP/3 issue continues to affect R2 custom domains; along with other active cases, it makes for procedural weather. Several status sources remain unavailable for quality checks. Market mood is Greed, which has filed no evidence. Filed and adjourned.

Latest readings

19:55Pinecone recovered: 5xx errors on some control plane operations resolved after 4h 50m
19:50CockroachDB Cloud recovered: Cloud Latency, Delays, Connection Issues resolved after 4h 13m
19:40Redis Cloud recovered: GCP US-Central1 issue resolved after 3h 39m
19:15Google Cloud recovered: Multiple products in us-central1-b are experiencing network service degradation. resolved after 4h 31m
19:15Clerk recovered: GCP incident in us-central1-b resolved after 3h 26m
19:10Twilio incident resolved: On Call Engineers are Investigating resolved; 8 degraded incidents remain
19:00Google Cloud degraded: Multiple products in us-central1-b are experiencing network service degradation.
18:45Cursor recovered: Degraded performance of Grok 4.5 resolved after 2h 2m

Warnings in effect

Hold fresh Vercel deployments; the weather has already submitted enough failure reports.

Advice for the day

Pause Vercel deployments.
Check affected Fireworks AI model workflows before relying on them.
Treat unavailable source checks as a missing instrument, not a clear forecast.