The migration of our Europe (eu1) cluster is now complete. Most gateways reconnected successfully.
Traffic for gateways that use the UDP forwarder are are configured with the IP address of our old server (instead of the domain name) is temporarily routed through Packet Broker while we reach out to customers that need to update their gateways.
Some gateways use a version of the UDP forwarder that starts separate uplink and downlink streams. In some situations these streams end up on different servers of our cluster, which causes issues with clock synchronization between our cluster and the gateway. As a result, downlinks may fail. We're working on a solution for this.
Posted Jul 20, 2020 - 12:28 CEST
Verifying
We have completed all steps to migrate our Europe (eu1) cluster to new infrastructure. Our North America (nam1), Australia (au1) clusters and the Global Join Server are operational again.
We're monitoring as gateways are reconnecting, and LoRaWAN traffic in our Europe (eu1) cluster is recovering.
Posted Jul 20, 2020 - 11:30 CEST
In progress
Scheduled maintenance is currently in progress. We will provide updates as necessary.
Posted Jul 20, 2020 - 11:00 CEST
Update
We would like to clarify that the maintenance window is between 09:00-10:30 UTC (11:00-12:30 CEST). A previous version of this announcement mentioned an incorrect time.
Posted Jul 20, 2020 - 08:50 CEST
Scheduled
Services on The Things Enterprise Stack - Cloud Hosted (V3) will be unavailable for up to 90 minutes. We are migrating our services to a new hosting infrastructure to improve availability, reliability and scalability. The migration allows us to do rolling updates, so customers won’t be experiencing downtime during future updates.
Any V2 services you may be using remain intact.
During this maintenance period, your account will be unavailable, and messages from your devices won't be routed to any application. Access will resume as soon as maintenance ends for which we will send you a notification via this channel.
A small number of UDP gateways on the `eu1` cluster may not automatically reconnect after migration. For these cases we have a temporary workaround in place that nevertheless makes sure traffic is routed to your application. We will notify the admins of any impacted gateways directly.
If you have questions or concerns about the impact of this maintenance, or if you experience anything unusual after completion, please reach out to support@thethingsindustries.com.
—
The Things Industries Team
Posted Jul 13, 2020 - 11:10 CEST
This scheduled maintenance affected: The Things Stack Cloud (Europe 1 (eu1), North America 1 (nam1), Australia 1 (au1), Join Server).