ingestr v1.0.55 - Vitess TLS and Postgres CDC Arrays

ingestr v1.0.55 was published on June 30, 2026, with release notes focused on CDC reliability for Vitess and Postgres. The main change is TLS support for the Vitess CDC VStream connection, which matters when a pipeline reads change events across secured database links. The same release also preserves Postgres CDC array columns and makes BigQuery load errors visible instead of leaving operators with a quiet failure path.

The full release notes and downloads are on the GitHub release page.

TLS for Vitess CDC VStream

The clearest operator change in v1.0.55 is TLS support for the Vitess CDC VStream connection. The changelog names this under mysql/vitess, so the scope is the Vitess CDC path rather than every MySQL style source in ingestr.

That matters for teams running Vitess with protected database links. A CDC reader that cannot speak TLS can force awkward network exceptions, sidecar workarounds, or separate access paths. This release closes that gap for the VStream connection itself.

The notes do not name a CLI flag, environment variable, or config key. That is worth saying plainly. Before changing a production ingestion job, operators should read the upstream release page and test the Vitess connection path in the same network shape used by the scheduler or worker.

Postgres CDC preserves array columns

The Postgres CDC fix is narrow and useful: postgres-cdc now preserves array columns. For data engineers, that is a schema fidelity issue more than a cosmetic one.

Array fields show up in event tables, application state, and denormalized source models. If the CDC path drops or distorts them, the downstream table can look valid while still losing source detail. That kind of problem often appears later, when a model expects a repeated value and gets an empty field or a changed shape.

The release note does not describe a migration step or a breaking change. The practical check is still simple: pick a source table with array columns, run a CDC sync in a test target, and compare the loaded shape against the source. If the target is BigQuery, pay attention to how repeated values or nested fields are represented after load.

BigQuery load errors are now visible

The other half of the Postgres CDC fix is BigQuery error visibility. The changelog says the release surfaces BigQuery load errors, which is a direct improvement for pipeline diagnosis.

Silent or poorly exposed load failures waste time because the extraction side can look healthy while the destination rejected data. BigQuery can reject loads for schema shape, type mismatch, bad records, or destination constraints. The release note does not list exact error classes, so the safe reading is that ingestr now gives operators a clearer failure signal when BigQuery load work fails.

This is especially relevant when paired with the array column fix. Array preservation changes what reaches the destination. Better load error reporting gives teams a faster way to see whether the destination accepted that shape.

Small release with data path impact

The changelog for v1.0.55 is short. It contains two merge entries and two substantive changes: Vitess VStream TLS support and a Postgres CDC fix that includes array preservation plus BigQuery load error reporting.

That is not a broad connector release. It is a focused data path release. The useful read for platform teams is whether they run Vitess CDC over secured links, whether they rely on Postgres array columns, or whether BigQuery load failures have been hard to trace in their ingestr jobs.

There is no prerelease flag on this release. There are also no migration notes in the published text. Treat it as a targeted update, then verify the CDC and destination behavior that matches your own connector mix.

Where to get it