v0.1.0

Initial release.

• Data source: CrateDB over the PostgreSQL wire protocol (pgx), with trust/password auth,
  TLS (disable/require/verify-ca/verify-full, inline PEM material or server-side
  certificate file paths), connection pool tuning with the PostgreSQL data source's defaults
  (100 open, 100 idle, 4-hour lifetime), and secure SOCKS proxy support. The config
  screen mirrors the built-in PostgreSQL data source's layout (Connection / Authentication /
  TLS/SSL Auth Details / Additional settings) so postgres users feel at home.
• Time-series macros: $__timeFilter (millisecond precision), $__dateFilter,
  $__timeFrom/$__timeTo, $__fromTime/$__toTime, $__time, $__timeEpoch,
  $__timeGroup(Alias) (DATE_BIN-based, with an optional fill argument: NULL, a number, or
  previous fills the buckets a query returned no rows for), $__unixEpochFilter,
  $__unixEpochFrom/To, $__unixEpochGroup(Alias), the nanosecond $__unixEpochNano* forms,
  $__interval_s, and $__conditionalAll. A numeric column named time in a time-series query
  becomes the time axis, its unit read from its magnitude as in the PostgreSQL data source, and
  a label column beside it splits the result into one series per label. Most macros resolve
  backend-side, so alerting works identically.
• Visual query builder: new queries open in a builder with Table, Time series, and Logs
  flavors — schema/table pickers backed by information_schema, typed filter rows, aggregations
  with grouping, order/limit. Picking a table is enough to produce the recommended
  $__timeGroupAlias/$__timeFilter aggregation: the time column (and, for logs, the
  message/level columns) is guessed from column metadata. Builder ⇄ SQL switching is two-way:
  the builder state rides along the query, and hand-written SQL is parsed back into builder
  state when it fits the builder's model — anything it can't represent asks before being
  replaced.
• Query editor: SQL editor with schema/table/column autocomplete (CrateDB
  information_schema, including sys), macro completion, hover docs on macros, a cheat
  sheet, and Ctrl/Cmd+Enter to run. The result
  format defaults to Auto, which infers time series vs. table from the query shape and
  shows what it resolved to; explicit overrides remain available. A $__timeGroup(...)
  used as a bare projection is shorthand for the aliased form, and EXPLAIN queries
  render as tables.
• Query guidance: an info notice on panels whose query has no time-range macro (no
  partition pruning; sys tables are exempt), a fallback message when autocomplete
  introspection fails, and a config-page warning when the host URL points at CrateDB's HTTP
  port instead of the PostgreSQL port.
• Template variables & ad-hoc filters: query variables use the same SQL editor as panels
  (autocomplete + macro hover, via CustomVariableSupport) rather than a bare text box, and
  resolve as value or text/value (__text/__value) pairs;
  multi-select via $__conditionalAll, and dashboard-wide ad-hoc filters applied to queries
  that read one plain table. Filter keys skip
  column types that can't back an equality filter (OBJECT, GEO, arrays), and a
  cratedb_adhoc_tables dashboard variable narrows the key picker on large schemas.
• Type mapping: CrateDB OBJECT columns surface as structured JSON fields, which a table
  panel's cell inspector opens in full; arrays keep their PostgreSQL text form.
• Logs: a Logs query format renders rows as log lines (e.g. in Explore); alias columns
  as time, body, and optionally level (see the cheat sheet's logs template). Explore's
  logs-volume histogram runs as a full-range aggregation (severity-bucketed when a level
  column is present) instead of bucketing only the returned lines — for hand-written SQL too,
  via the SQL-to-builder parser.
• Bundled dashboards: CrateDB Cluster Health (sys-schema monitoring, no separate
  exporter) and CrateDB Getting Started (the recommended query pattern plus an OBJECT
  column, a Logs-format panel, and an annotation query — all runnable against the dev
  stack's seeded demo tables).
• Turnkey demo stack: docker compose up on a clone builds the plugin, starts CrateDB,
  seeds the demo tables, and serves Grafana with the data source provisioned and both bundled
  dashboards populated. Docker is the only host requirement.
• Cluster protection: configurable row limit and a TTL cache for autocomplete
  introspection.
• Diagnostics: actionable connection error messages (auth / TLS / network / timeout) on
  the config page and downstream error attribution for queries.
• Documentation: provisioning reference, annotation column convention, a migration guide
  from the PostgreSQL data source, the design document, and a roadmap.