Minor Changes

• d91e66b8: Make the external API v2 rate limit (/api/v2/*) configurable via EXTERNAL_API_RATE_LIMIT_MAX. Defaults to 100 requests/minute, matching the previous hardcoded value.

Patch Changes

• 1e355ac4: fix: don't scan the whole table to discover Map keys

  getMapKeys only applied a time predicate when the caller passed both a date range and a timestamp expression; otherwise the raw sampledKeys scan ran with no WHERE and touched every part of the table. It now defaults a missing date range to the last 24 hours and skips the raw scan entirely when there is no timestamp expression to filter on. The chart, alert and dashboard-filter editors pass the source and date range they already have so Map keys keep autocompleting there.

• 793fe19e: feat: Substitute dashboard variables in markdown tiles

• ec4f5087: fix: allow grouping gauge and sum metric charts by materialized and alias columns

  Grouping or selecting a MATERIALIZED or ALIAS column on a gauge or sum metric chart failed with Unknown expression identifier, because the intermediate query didn't carry those columns through.

• 08d9a908: feat: Plot several PromQL expressions on one chart

• 19182e7e: fix: only split a query at a standalone SETTINGS keyword

  extractSettingsClauseFromEnd cut the query at the first "settings" anywhere in it, including inside a string or an identifier. A multi-series metric chart with a metric such as app.settings.reloads produced SQL with an unterminated string, and a SQL filter on a column such as AppSettings was rewritten to reference App. The keyword now has to stand alone outside quotes.

• d9e2c58b: Add a per-source floor for auto granularity. A metric source can now set "Minimum auto granularity" (Team Settings → Sources → your Metrics source) so that auto-inferred time buckets never go below it — useful when the underlying metric is reported on a fixed interval (e.g. a 60s scrape), since a short selected date range can otherwise auto-infer a smaller bucket than that interval and render a sparse/steppy series (alternating real-sample/empty buckets). Mirrors Grafana's per-datasource "Min interval" setting. Unset (the default) preserves the existing unfloored behavior, and an explicit (non-auto) granularity chosen on a tile is never affected.

• a5e7841c: fix(mcp): time-bound the trace waterfall span/log fetches so they prune partitions instead of scanning the full retention window. The clickstack_trace_waterfall tool now threads the search window into its span and correlated-log queries, adds a max_execution_time ceiling, and probes a trace's [min, max] span extent so an explicit traceId older than the default window — or a trace that ran longer than an hour — still resolves in full. The probe also runs for auto-picked traces, so a picked trace whose root predates the window is no longer truncated into a partial tree. The fetch window is width-clamped to the recent tail so a reused or sentinel traceId (e.g. an all-zero id from an uninstrumented emitter) can't widen the scan back toward the retention edge or stitch unrelated occurrences into one tree. Empty-result hints now name the recoverable action (pass an explicit startTime) instead of the misleading "widen the window", and a probe that fails is surfaced rather than silently swallowed. ClickHouse query timeouts are also reclassified from user to server errors.

• Updated dependencies [1e355ac4]

• Updated dependencies [793fe19e]

• Updated dependencies [ec4f5087]

• Updated dependencies [8a9fcd2f]

• Updated dependencies [08d9a908]

• Updated dependencies [a8a72c11]

• Updated dependencies [8cfb2672]

• Updated dependencies [4570d5ee]

• Updated dependencies [78ed5921]

• Updated dependencies [23423c13]

• Updated dependencies [19182e7e]

• Updated dependencies [d9e2c58b]

  • @hyperdx/common-utils@0.30.0