Mastra ClickHouse adapter targets observability scale

This is the right kind of boring infrastructure upgrade: it doesn’t change what Mastra does, but it makes the observability stack credible for real production workloads.

• –ClickHouse is a better fit than row-oriented storage when the primary workload is time-range filtering, rollups, and dashboard aggregations
• –Mastra is clearly separating “debug one run” data from “analyze the fleet” data, which is the right mental model for agent ops
• –The new `ObservabilityStorageClickhouseVNext` path and `@mastra/core@1.29.0` requirement suggest this is the production-grade schema, not a cosmetic adapter
• –This matters most for teams with ephemeral infrastructure or growing telemetry volume, where local files or lighter stores start to fall over
• –Studio’s value increases here because it can pair aggregate metrics with per-run traces without asking teams to stitch together their own analytics layer