Operations / OPS-33
Textbridge outage: SMS codes to UK and Irish numbers delayed for about two hours
Description
Summary
Textbridge, our SMS provider, had a fault on its route to UK and Irish mobile networks. From 07.05 to 09.10 phone codes to UK and Irish numbers arrived up to two hours late, or after they had expired (codes work for ten minutes). Numbers elsewhere, passwords and single sign-on were not affected. Summary for customers: Confluence OPS › OPS-33: SMS codes delayed (Textbridge outage).
Incident lead: Elif Yildiz, with Piet de Vries.
Impact
- Northgate: site workers on phone codes at 9 of its 12 sites, during the early shift.
- Greenhollow Dairy: workers at all 6 sites.
- About 260 people tried to sign in; about 430 code requests failed or arrived late.
- Support's ticket for all customers: SUP-963 (the customers' own tickets are linked from it).
Timeline (Wednesday 17 June, CEST)
- 07.05 Codes to +44 and +353 numbers stop arriving. Our side shows them accepted by Textbridge.
- 07.40 A Northgate site manager calls support: nobody at the gate can sign in.
- 07.55 Elif sees delivery receipts for UK numbers 40 minutes behind. No alarm had fired: we only watched "accepted".
- 08.10 Textbridge confirms a fault on its UK and Irish route.
- 08.20 Support emails Northgate's and Greenhollow's administrators: workers already signed in are fine; others use the site's paper sign-in for the morning.
- 09.10 Receipts back to normal. Late codes have expired; people ask for new ones.
- 09.40 Support confirms with both customers that sign-in works.
Why it hurt more than it had to
- We heard about it from a customer, 35 minutes in. We measured whether Textbridge accepted a code, not whether it was delivered.
- We had nowhere to tell customers at once.
- One SMS provider, no fallback.
Follow-ups
- Alarm when receipts lag more than 5 minutes: ENG-1496
- Status page: ENG-1497, OPS-34
- Phone-code screen says when codes are late: ENG-1505
- An escalation contact at Textbridge, in the shared vault
- A second SMS provider: an idea for product
Linked issues
- relates to ENG-1488 Page on-call when refused or undelivered texts pass 5% in 10 minutes Closed
- relates to ENG-1496 Page on-call when SMS delivery receipts lag more than five minutes Done
- relates to ENG-1505 Phone-code screen says when codes are arriving late Done
- relates to ENG-1497 Status page and incident updates Done
- relates to SUP-963 Phone codes to UK and Irish numbers not arriving (Textbridge) Done
- relates to ENG-1509 Administrators can give a person a one-time code when phone codes are late Done
Comments (2)
Oskar Lund · 22 Jun 2026, 13:40 UTC
Review done in Emargen with Elif, Piet and Ines, the first from the new Incident review template. Follow-ups linked.
Textbridge's report: a configuration change on their UK and Irish gateway sent messages to a dead route; their monitoring watched acceptance, like ours. Escalation contact is in the shared vault.