CleverCloud incident

Network Groups: redeployed instances not joining their network group

Resolved after 1 day 4 hours, affecting NetworkGroups, on November 20, 2026.

investigating Nov 20, 2026, 02:16 PM UTC

Between November 20, 15:41 and 19:14 UTC, instances started or redeployed on Clever Cloud were not added to their Network Group, so they could not be reached through it. Instances that were already running were not affected. The issue started degrading from 15:16 UTC. Cause: an internal component of the Network Group service ran into an inconsistent state linked to a single organisation. Instead of isolating it, the service stopped processing membership updates for all customers. Resolution: we identified the cause at 17:16 UTC and corrected the inconsistent state at 19:12 UTC. Pending membership updates were applied by 19:14 UTC. Our support team resynchronised the customers who were still affected. Follow-up: since November 20, 23:25 UTC, an error affecting one organisation can no longer block membership updates for others. That organisation is set aside and retried on its own. We also added monitoring that alerts our team if membership updates stall, and we checked every organisation that uses Network Groups: no other one was in this state. If one of your instances is still not reachable through its Network Group, redeploy it or contact support.

Started
Nov 20, 2026, 02:16 PM UTC
Resolved
Nov 21, 2026, 07:11 PM UTC
Duration
1d 4h
Detected by Pingoru
Nov 20, 2026, 02:16 PM UTC

Affected components

NetworkGroups

Update timeline

  1. investigating Nov 20, 2026, 02:16 PM UTC

    Between November 20, 15:41 and 19:14 UTC, instances started or redeployed on Clever Cloud were not added to their Network Group, so they could not be reached through it. Instances that were already running were not affected. The issue started degrading from 15:16 UTC. Cause: an internal component of the Network Group service ran into an inconsistent state linked to a single organisation. Instead of isolating it, the service stopped processing membership updates for all customers. Resolution: we identified the cause at 17:16 UTC and corrected the inconsistent state at 19:12 UTC. Pending membership updates were applied by 19:14 UTC. Our support team resynchronised the customers who were still affected. Follow-up: since November 20, 23:25 UTC, an error affecting one organisation can no longer block membership updates for others. That organisation is set aside and retried on its own. We also added monitoring that alerts our team if membership updates stall, and we checked every organisation that uses Network Groups: no other one was in this state. If one of your instances is still not reachable through its Network Group, redeploy it or contact support.

Your product runs on more than CleverCloud

Add the vendors you depend on and get one message when any of them reports trouble. Free for 3 vendors, no card.

Add your stack →