Current availability for Relay Cloud, runners, cache, and related platform services.
All systems operational
Last updated Jul 13, 2026, 9:25 PM UTC
90-day uptime
Relay Cloud API
REST API, webhooks, and pipeline triggers
Pipeline execution
Job scheduling, queues, and orchestration
Cloud runners
Linux, macOS, and Windows managed runners
Self-hosted runners
Agent registration and job routing
Remote cache
Cross-branch artifact and dependency caching
Preview environments
PR preview deploys and teardown
Dashboard & CLI auth
Sign-in, SSO, and token issuance
Relay AI
Flaky test detection and pipeline insights
A misconfigured cache shard caused slower restores for large monorepo jobs in us-east-1. Traffic was shifted to healthy nodes and latency returned to baseline.
Duration: 47 minutes
A surge in matrix builds increased macOS queue depth. Additional runner capacity was provisioned and queue times normalized.
Duration: 1 hour 12 minutes
Webhook workers fell behind during a traffic spike. Backlog was cleared and retry policies prevented duplicate deliveries.
Duration: 23 minutes
Current availability for Relay Cloud, runners, cache, and related platform services.
All systems operational
Last updated Jul 13, 2026, 9:25 PM UTC
90-day uptime
Relay Cloud API
REST API, webhooks, and pipeline triggers
Pipeline execution
Job scheduling, queues, and orchestration
Cloud runners
Linux, macOS, and Windows managed runners
Self-hosted runners
Agent registration and job routing
Remote cache
Cross-branch artifact and dependency caching
Preview environments
PR preview deploys and teardown
Dashboard & CLI auth
Sign-in, SSO, and token issuance
Relay AI
Flaky test detection and pipeline insights
A misconfigured cache shard caused slower restores for large monorepo jobs in us-east-1. Traffic was shifted to healthy nodes and latency returned to baseline.
Duration: 47 minutes
A surge in matrix builds increased macOS queue depth. Additional runner capacity was provisioned and queue times normalized.
Duration: 1 hour 12 minutes
Webhook workers fell behind during a traffic spike. Backlog was cleared and retry policies prevented duplicate deliveries.
Duration: 23 minutes