Is IBM Cloud Down? Live status & outage reports

Service info and troubleshooting

Console Won't Load or Freezes

The IBM Cloud console is a heavy single-page app — if it hangs or shows a blank screen, start with a hard refresh (Ctrl+Shift+R). Clear the browser cache and switch to Chrome or Firefox if you're on Safari or Edge. Disable browser extensions, especially ad blockers, which often intercept API calls the console depends on.

API Returns 401 or 403 Errors

This almost always means the IAM token has expired or the API key lacks the right permissions. Regenerate the API key in Access (IAM) → API keys, and double-check that the service ID or user account has the correct role assigned to the target resource group. A 403 with a valid key usually points to a missing policy — check Access policies under your account.

Resource Provisioning Stuck or Failed

• Open the Activity Tracker or Notifications panel to see the actual error code behind the stuck status.
• If provisioning sits in 'in progress' for more than 15 minutes, try deleting and recreating the resource in a different region.
• Quota limits are a common silent blocker — check Account → Quotas before retrying.
• If the error mentions 'ETCD' or 'backend timeout,' it's a platform-side hiccup; wait 10–15 minutes and retry.

Object Storage Files Won't Upload

Large file uploads to Cloud Object Storage often fail because of client-side timeouts. Use the IBM COS SDK or the ibmcloud cos CLI instead of the web UI for anything over 500 MB. Enable multipart upload and set a reasonable part size (100–200 MB). Also verify that the bucket's CORS policy allows your origin if you're uploading from a browser app.

SSH Access to Virtual Server Times Out

• Confirm the security group attached to the instance allows inbound TCP on port 22 from your IP.
• Check that the floating IP is correctly associated — a reboot sometimes detaches it.
• If the instance is reachable but SSH still hangs, connect via the VNC console in the portal and verify that the sshd service is running.

High Latency on Kubernetes Workloads

Spikes in pod response time usually trace back to node resource saturation. Run kubectl top nodes and kubectl top pods to spot the culprit. If CPU and memory look fine, check network policy rules — overly broad deny-all policies can force traffic through unexpected routes and add tens of milliseconds per hop. Also review whether your pods are co-located in a single zone; spreading them across availability zones with proper topology keys usually smooths things out.