From StreamBuild to SQLTide

SQLTide 0.60.0 is the renamed package. The command is sqt, not stb. There is no compatibility alias for authored configuration or Python imports.

Change the authored names

• Install sqltide in place of streambuild in every development, server and CI environment.
• Rename streambuild_project.toml → sqltide_project.toml and streambuild_local.toml → sqltide_local.toml.
• Change STREAMBUILD_* environment variables to SQLTIDE_*.
• Import sqltide instead of streambuild in macros, sensors and providers.
• Change trusted proxies to the default X-SQLTide-User header, or explicitly configure your header.
• Ignore .sqltide/ in place of .streambuild/. Browser sessions must sign in again.

Let SQLTide migrate stored state

Quiesce writers and upgrade all machines and CI jobs that write to a target together. The first connected sqt command renames _streambuild_* metadata to _sqltide_*; it refuses while an old StreamBuild run still holds the target lock. Do not remove a lock until you have confirmed no warehouse mutation is still running. StreamBuild cannot read a target after it has been renamed.

Accounts in .streambuild/ or a shared control-store database are migrated too. Managed Kafka sources keep their existing consumer groups and committed offsets. Version 0.60.0 still generates streambuild_ group defaults when no group is specified. Preserve an existing live identity: changing its prefix creates a different consumer group and does not transfer committed offsets. The new-project examples explicitly configure sqltide_ groups; do not copy those values over an existing source. Do not reset broker offsets or drop metadata tables as part of this upgrade.

Verify before rebuilding

sqt --version
sqt discover
sqt compile
sqt plan

Inspect the connected plan before approving a build. Check source groups, target database, retained history, role assignments and proxy authentication. A successful compile alone does not prove the warehouse or access configuration is correct.