When sending fails

Work from the outside in. Stop at the first thing that is wrong.

One customer only

1. Credit. Does the bind or account have balance?
2. Bind. Is the bind enabled? Is it connected in Live status?
3. IP. Is the customer connecting from an IP on their list? Connections from other IPs are refused at the firewall and never reach SMPP.
4. Country. Is the destination country on the bind?
5. Speed. Are they getting ESME_RTHROTTLED? They send faster than their limit.
6. Sender ID. Is it allowed for that country?
7. HTTP API. Is the number a Malawi number? Other countries get destination_not_allowed.

Everyone

1. Upstream bind. Is it connected? See Upstream bind.
2. Upstream credit. Is the upstream account balance above zero?
3. Services. Are all containers running and healthy? Check with docker compose ps in the yanu-infra folder on the server.
4. Disk. Is the disk full? A full disk stops the database and the queues.
5. Database. Can the services reach PostgreSQL?
6. Recent deploy. Did something change in the last hour? Roll back by reverting the commit on main.

Messages accepted but not delivered

• Look at receipts: many UNDELIV or EXPIRED for one network point to a network problem.
• Compare with a test SMS to your own phone.
• Report the pattern, times and sample message ids to the upstream provider.

Tell customers

If sending is down for everyone, tell customers early, by email from [email protected] and by phone for big accounts. Say what is affected, and when you will give the next update.