Cloud Run executes containers serverlessly and scales them to zero - which means it inherits the health of everything underneath it: the regional serving infrastructure, container registry reads, and the build path if you deploy from source. Incidents usually show up as elevated cold-start latency, request 5xx in one region, or deploys failing while running revisions keep serving.
Elevated cold-start latency after capacity or scheduler events
Request 5xx rates in one region while other regions stay healthy
New deploys or revisions failing (image pulls, build path) while the serving revision is fine
Event-driven invocations lagging because Pub/Sub or Eventarc delivery is delayed upstream
Cloud Run questions
Is Cloud Run down right now?+
The live status above reflects our latest check. Cloud Run failures are almost always regional - if you serve from a single region, that region’s health is your availability. Multi-region behind a global load balancer is the standard fix.
Why did my Cloud Run latency spike during a GCP incident?+
Cold starts amplify upstream incidents: if instances are recycled or scaled during an event, every new instance pays image-pull and startup cost simultaneously. Setting min-instances above zero for latency-sensitive services buys you through most of it.
Does Cloud Run have an SLA?+
Yes - 99.95% monthly uptime for requests served. Tiered credits apply below target, claimed within 30 days, capped at 50% of the month’s Cloud Run charges.
Get alerted when Cloud Run breaks
Instant email the moment we detect an incident - affected services and regions included.
gcp