Upgrade to The Things Network v3.16.0 (completed)

TTN Announcements
1

Completed: Upgrade to The Things Network v3.16.0

This is a cross-post of an incident on our status page. It will be updated automatically.

Impact: Maintenance

Scheduled: Mon, 15 Nov 2021 09:30:00 +0100 until 11:30:00 +0100

Resolved: Mon, 15 Nov 2021 11:30:29 +0100

Affected Components

  • Europe 1 (eu1.cloud.thethings.network): Operational
  • North America 1 (nam1.cloud.thethings.network): Operational
  • Australia 1 (au1.cloud.thethings.network): Operational

Scheduled

Posted: Fri, 12 Nov 2021 16:07:45 +0100

During this maintenance window we will upgrade The Things Network v3.16.0

We do not expect noticeable downtime during this deployment.

Here is the changelog since the current version 3.15.2:

Added

  • ttn_lw_as_subscription_sets_publish_success_total and ttn_lw_as_subscription_sets_publish_failed_total metrics to track the number of subscription set publish attempts.
  • Application Server advanced distribution settings:
    • as.distribution.global.individual.subscription-blocks controls if the Application Server should block while publishing traffic to individual global subscribers (such as MQTT clients).
    • as.distribution.global.individual.subscription-queue-size controls how many uplinks the Application Server should buffer for an individual global subscriber. Note that when the buffer is full, the Application Server will drop the uplinks if --as.distribution.global.individual.subscription-blocks is not enabled. Use a negative value in order to disable the queue.
    • as.distribution.local.broadcast.subscription-blocks controls if the Application Server should block while publishing traffic to broadcast local subscribers (such as webhooks and application packages matching).
    • as.distribution.local.broadcast.subscription-queue-size controls how many uplinks the Application Server should buffer for an broadcast local subscriber. Has the same semantics as --as.distribution.global.individual.subscription-queue-size.
    • as.distribution.local.individual.subscription-blocks controls if the Application Server should block while publishing traffic to individual local subscribers (such as PubSub integrations).
    • as.distribution.local.individual.subscription-queue-size controls how many uplinks the Application Server should buffer for an individual local subscriber. Has the same semantics as --as.distribution.global.individual.subscription-queue-size.
  • ttn_lw_gs_txack_received_total, ttn_lw_gs_txack_forwarded_total and ttn_lw_gs_txack_dropped_total metrics, which track the transmission acknowledgements from gateways.
  • gs.txack.receive, gs.txack.drop and gs.txack.forward events, which track the transmission acknowledgements from gateways.
  • Remote federated authentication providers and external user management.
    • Authentication providers can be managed via the AuthenticationProviderRegistry service. See the tti-lw-cli authentication-providers commands for CLI support.
    • External users can be managed via the ExternalUserRegistry service. See the tti-lw-cli external-users commands for CLI support.
  • External users (users associated with a federated authentication provider) may no longer update their passwords, or request temporary passwords.
  • External users will have their information (name and email) updated on login, based on the data provided by the authentication provider.
  • Users may be associated to multiple authentication providers at the same time.
    • This requires a database schema migration (ttn-lw-stack is-db migrate) due to the relaxed unique constraint.
  • ttn-lw-stack as-db migrate command to migrate the Application Server database. This command records the schema version and only performs migrations if on a newer version.
    • Use the --force flag to force perform migrations.
  • Server-side event filtering with the names field.

Changed

  • Gateway Server default UDP worker count has been increased to 1024, from 16.
  • Application Server webhooks and application packages default worker count has been increased to 1024, from 16.
  • Application Server no longer sets the end device's session.started_at and pending_session.started_at. The session start time should be retrieved from the Network Server, per API specification.
    • This requires an Application Server database migration (ttn-lw-stack as-db migrate) to clear the started_at field in existing (pending) sessions.
  • Console changing to server-side event filtering (used to be client-side).

Removed

  • The ttn_lw_gs_status_failed_total, ttn_lw_gs_uplink_failed_total metrics. ttn_lw_gs_status_dropped_total and ttn_lw_gs_uplink_dropped_total should be used instead, as they contain the failure cause.
  • The gs.status.fail and gs.up.fail events. gs.status.drop and gs.up.drop should be used instead, as they contain the failure cause.
  • The data_rate_index field in uplink message metadata. Observe the fully described data rate in the data_rate field instead.
  • LoRaWAN data rate index reported to LoRa Cloud DMS.
  • Dockerfile doesn't define environmental variables TTN_LW_BLOB_LOCAL_DIRECTORY, TTN_LW_IS_DATABASE_URI and TTN_LW_REDIS_ADDRESS anymore. They need to be set when running the container: please refer to docker-compose.yml for example values.
  • CockroachDB from development tooling as well as config option within docker-compose.yml.
    • This also changes the default value of the --is.database-uri option, so it can connect to the development Postgres database by default.

Fixed

  • LoRaWAN Backend Interfaces AppSKeyReq and HomeNSReq interop calls in multi-tenant deployments.
  • Handling of NaN values in our JSON API.
  • Receiver metadata from more than one antenna is now available in messages received from Packet Broker.
  • Unhelpful error message when aborting the OIDC Login in the Console.
  • Parsing of multi-word description search queries.

Scheduled

Posted: Fri, 12 Nov 2021 16:15:27 +0100

We will be undergoing scheduled maintenance during this time.

In Progress

Posted: Mon, 15 Nov 2021 09:30:32 +0100

Scheduled maintenance is currently in progress. We will provide updates as necessary.

Completed

Posted: Mon, 15 Nov 2021 11:30:29 +0100

The scheduled maintenance has been completed.

1 Like
2

The incident on our status page was just updated with new information. The first post in this topic has been updated accordingly.

3

The incident on our status page was just updated with new information. The first post in this topic has been updated accordingly.