Hector Hector

Every Postgres and MySQL service in the catalog gained a SQL tab with an editor, a schema browser, a history of your past queries, and a results table that handles thousands of rows without flinching. Cmd/Ctrl+Enter to run, a cancel button for queries that got too warm, drafts that survive a tab switch, and error messages that speak your language. Every execution is read-only by construction, so you can explore freely without fear of breaking production.

Two new catalog services bring native change-data-capture (CDC) to your project: Debezium for the heavy cases (Postgres, MySQL, MongoDB → NATS, Kafka, HTTP), and Conduit for lighter pipelines between varied sources and sinks. The wizard recognises the other services in your project, so you just pick the source and the destination from a dropdown, credentials and endpoints are on us. Use it to feed caches, react to changes in real time, or pipe data out of Guara Cloud without having to reinvent the pipeline.

Four new observability surfaces shipped to the CLI: guara services metrics for metrics and ASCII charts, guara services traces for listing and drawing traces as a waterfall, guara projects map for your project’s service graph, and guara projects observatory for a live panel with health, events, and traces. Everything accepts --watch for in-place refresh and --json for piping into jq. In the same wave, guara catalog query and guara catalog schema bring SQL into the terminal, perfect for pasting into a Makefile or a migration script.

We rebuilt the platform’s telemetry layer so every log line and every trace span already ships with rich context, who acted, in which project, in which service, in which environment. That context follows the request all the way through, even across internal queues and async jobs. For you, this means: when something goes wrong, support opens your case and finds your needle in seconds, and we can spot regressions before you feel them. It’s the kind of improvement you don’t see, you just notice that questions become easier to answer.

A new Audit Logs page arrived under Operations, dedicated to your own actions. Filter by action type, resource, project, and time window. Every row expands to show full metadata, who, from where, with what outcome. The view is strictly yours: you never see what other accounts did, and nobody else sees what you did. The page is fully keyboard-navigable and works on small screens, so you can audit a suspicious session from your phone if you need to.

The new SQL tab is hardened against the classic injection and statement-smuggling tricks, and execution limits are now consistent even when the platform scales to multiple nodes serving the same team. The new CDC services come from verified releases with integrity checks at boot, so what you deploy is what we shipped. Error messages across the dashboard got clearer and never expose server internals; on the flip side, standardised error codes help the CLI and the API suggest the right next step. And the platform dashboard gained volume-usage panels, so nobody gets surprised by a full disk at three in the morning.