github-actions · Nov 13, 2026 · Nov 13, 2026
@@ -9,8 +9,8 @@ NEXT_ALL_IN_ONE_IMAGE_NAME_DOCKERHUB=clickhouse/clickstack-all-in-one
 ALL_IN_ONE_IMAGE_NAME_DOCKERHUB=hyperdx/hyperdx-all-in-one
 NEXT_OTEL_COLLECTOR_IMAGE_NAME_DOCKERHUB=clickhouse/clickstack-otel-collector
 OTEL_COLLECTOR_IMAGE_NAME_DOCKERHUB=hyperdx/hyperdx-otel-collector
-CODE_VERSION=2.44.0
-IMAGE_VERSION_SUB_TAG=.44.0
+CODE_VERSION=2.45.0
+IMAGE_VERSION_SUB_TAG=.45.0
 IMAGE_VERSION=2
 IMAGE_NIGHTLY_TAG=2-nightly
 IMAGE_LATEST_TAG=latest

@@ -6,6 +6,109 @@ PR — keep the `hyperdx-release-notes` comment marker intact when editing so yo
 edits survive regeneration. Per-package detail lives in each
 `packages/*/CHANGELOG.md`.
 
+## v2.45.0 — 2026-11-13
+
+<!-- hyperdx-release-notes version=2.45.0 inputs=8861d0a1698f -->
+
+**PromQL evaluation choice and ClickHouse 26.10**
+
+You can now choose how a PromQL source on ClickHouse is evaluated — through the
+Prometheus HTTP API or the `prometheusQuery` table functions — instead of
+leaving HyperDX to probe the server on every query. The service map can also now
+be split by up to three extra group-by expressions, so a single slow or failing
+pod no longer disappears into its service average. This release also moves the
+all-in-one image and the Docker Compose files to ClickHouse 26.10.3.7. Two
+things change underneath you: TimeSeries tables created from now on default to
+schema version 8, which keeps each series' time range in a separate inner table,
+and ClickHouse's own `system.metric_log` switches to the bucketed schema. Your
+existing TimeSeries tables keep the layout they already have, but metric_log
+rows written before the upgrade are left behind in `system.metric_log_0`, so
+check anything that reads that table before you upgrade.
+
+### 💥 Breaking Changes
+
+- **`system.metric_log` moves to the bucketed schema**: ClickHouse now writes
+  `system.metric_log` in its bucketed form, and the rows recorded before the
+  upgrade move to `system.metric_log_0`. A dashboard, alert or saved query
+  reading ClickHouse's own metrics from `system.metric_log` will see only
+  post-upgrade rows until you point it at both tables.
+- **New TimeSeries tables default to schema version 8**: TimeSeries tables
+  created after the upgrade keep each series' time range in a separate inner
+  table. Existing tables are untouched and keep their current layout, so a table
+  created after the upgrade no longer matches the shape of one created before
+  it.
+- **`clickstack_service_health` no longer returns a verdict**: the MCP tool has
+  stopped labelling services critical, degraded or healthy, and no longer tells
+  an agent a service needs "no action" — each service keeps its threshold flags,
+  score and a factual summary, so agents reason from the numbers rather than a
+  verdict that service-wide totals can get wrong. `summary.statusCounts` is
+  replaced by `summary.servicesFlagged`, and the span attributes
+  `mcp.service_health.critical_count` and `mcp.service_health.degraded_count`
+  are replaced by `mcp.service_health.flagged_count`, so update anything that
+  reads those fields.
+
+### ✨ New Features
+
+- **Choose how a PromQL source is evaluated**: a PromQL source on ClickHouse can
+  now be set to use either the Prometheus HTTP API or the `prometheusQuery`
+  table functions, rather than having HyperDX work it out for itself. A source
+  pinned to the HTTP API against a server that does not serve it now fails with
+  a clear error instead of an obscure one, and onboarding pins a PromQL source
+  it detects to the HTTP API when the server serves it.
+- **Split the service map by extra dimensions**: you can now add up to three
+  extra group-by SQL expressions to the service map — for example
+  `ResourceAttributes['k8s.pod.name']` — so a node becomes a service plus those
+  values and a slow or failing pod stands on its own instead of being averaged
+  into its service.
+
+### 🔧 Improvements
+
+- **PromQL label lookups go through ClickHouse's Prometheus HTTP API**: on
+  ClickHouse 26.9+ with the `prometheus_api_v1` handler, label and label-value
+  lookups for a ClickHouse-backed PromQL source now proxy to the handler instead
+  of querying the TimeSeries tags table with SQL. Older servers, servers without
+  the handler, and sources set to `table-function` keep the SQL path, so a
+  source that runs queries over HTTP on 26.6-26.8 still gets label
+  autocomplete.
+- **The MCP `describe_source` tool reads its label catalogue the same way**:
+  asked about a ClickHouse-backed PromQL source, `describe_source` now lists
+  label and metric names through the `prometheus_api_v1` handler on ClickHouse
+  26.9+, so an agent sees the same catalogue the app does. It falls back to SQL
+  on older servers, servers without the handler, and sources set to
+  `table-function`.
+- **Fewer probes for sources without an explicit choice**: a PromQL source that
+  does not say how it should be evaluated is still probed, but the answer is now
+  kept for 5 minutes instead of being worked out again on every query.
+
+### 🐛 Bug Fixes
+
+- **`{{note}}` is filled in when an alert actually fires**: a webhook body using
+  `{{note}}` now carries the alert's note on real alert notifications, not only
+  on "Send test", so the context you wrote on the alert reaches the people
+  paged by it.
+- **MCP webhook and alert tools are declared open-world**: saving a webhook or
+  an alert sends HTTP requests to third-party endpoints, so
+  `clickstack_save_webhook` and `clickstack_save_alert` now carry
+  `openWorldHint: true`. MCP clients that treat open-world tools differently —
+  by asking you to confirm before they run, for instance — will now do so for
+  these two.
+
+### 📦 Build / Packaging
+
+- **Bundled ClickHouse is now 26.10.3.7**: the all-in-one image and the Docker
+  Compose files use `clickhouse/clickhouse-server:26.10.3.7-alpine`.
+
+<!-- hyperdx-package-list -->
+
+### 📦 Package changelogs
+
+- `@hyperdx/api` 2.44.0 → 2.45.0 — [changelog](https://github.com/ClickHouse/ClickStack/blob/main/packages/api/CHANGELOG.md#2450)
+- `@hyperdx/app` 2.44.0 → 2.45.0 — [changelog](https://github.com/ClickHouse/ClickStack/blob/main/packages/app/CHANGELOG.md#2450)
+- `@hyperdx/common-utils` 0.34.0 → 0.35.0 — [changelog](https://github.com/ClickHouse/ClickStack/blob/main/packages/common-utils/CHANGELOG.md#0350)
+- `@hyperdx/otel-collector` 2.44.0 → 2.45.0 — [changelog](https://github.com/ClickHouse/ClickStack/blob/main/packages/otel-collector/CHANGELOG.md#2450)
+
+<!-- /hyperdx-package-list -->
+
 ## v2.44.0 — 2026-11-11
 
 <!-- hyperdx-release-notes version=2.44.0 inputs=99e8264954db -->

@@ -1,5 +1,48 @@
 # @hyperdx/api
 
+## 2.45.0
+
+### Minor Changes
+
+- 2d17de5d7: feat: serve PromQL label lookups from ClickHouse's Prometheus HTTP API
+
+  On ClickHouse 26.9+ with the `prometheus_api_v1` handler, `/labels` and `/label/:name/values` for a ClickHouse-backed PromQL source now proxy to the handler instead of querying the TimeSeries tags table with SQL. Older servers, servers without the handler, and sources set to `table-function` keep the SQL path, so a source that runs queries over HTTP on 26.6-26.8 still gets label autocomplete.
+
+- db4e7ba67: chore: upgrade bundled ClickHouse to 26.10.3.7
+
+  The all-in-one image and the Docker Compose files now use
+  `clickhouse/clickhouse-server:26.10.3.7-alpine`. New TimeSeries tables default
+  to schema version 8, which keeps each series' time range in a separate inner
+  table; existing tables keep their layout. `system.metric_log` switches to the
+  bucketed schema, and rows from before the upgrade move to `system.metric_log_0`.
+
+### Patch Changes
+
+- c17fc16a2: fix: `{{note}}` in webhook bodies now contains the alert's note when the alert
+  fires, not only in "Send test"
+- 1fba1c381: fix: mark the MCP `clickstack_save_webhook` and `clickstack_save_alert` tools as open-world
+
+  Saved webhooks and alerts send HTTP requests to third-party endpoints, so both tools now declare `openWorldHint: true`.
+
+- 7b03880eb: feat: read the MCP PromQL label catalog from ClickHouse's Prometheus HTTP API
+
+  `describe_source` on a ClickHouse-backed PromQL source now lists label and metric names through the `prometheus_api_v1` handler on ClickHouse 26.9+, and falls back to SQL on older servers, servers without the handler, and sources set to `table-function`.
+
+- ac913ceb3: feat: Choose how a PromQL source on ClickHouse is evaluated (the Prometheus HTTP API or the `prometheusQuery` table functions) instead of probing the server on every query. Without a choice the server is still probed, but the answer is now kept for 5 minutes. A source set to the HTTP API against a server that does not serve it fails with a clear error. Onboarding pins a PromQL source it detects to the HTTP API when the server serves it.
+- 7cfd333ea: fix(mcp): drop the health verdict from clickstack_service_health
+
+  The tool no longer labels services critical, degraded or healthy, and no longer
+  tells agents a service needs "no action". Each service keeps its threshold
+  flags, score and a factual summary, so agents reason from the numbers instead of
+  a verdict that service-wide totals can get wrong. `summary.statusCounts` is
+  replaced by `summary.servicesFlagged`, and the span attributes
+  `mcp.service_health.critical_count` and `mcp.service_health.degraded_count` are
+  replaced by `mcp.service_health.flagged_count`.
+
+- Updated dependencies [2d17de5d7]
+- Updated dependencies [ac913ceb3]
+  - @hyperdx/common-utils@0.35.0
+
 ## 2.44.0
 
 ### Minor Changes

@@ -1,6 +1,6 @@
 {
   "name": "@hyperdx/api",
-  "version": "2.44.0",
+  "version": "2.45.0",
   "license": "MIT",
   "private": true,
   "engines": {
@@ -23,7 +23,7 @@
     "@clickhouse/clickstack-node": "^1.0.2",
     "@clickhouse/client-common": "1.23.0-head.fae5998.1",
     "@esm2cjs/p-queue": "^7.3.0",
-    "@hyperdx/common-utils": "^0.34.0",
+    "@hyperdx/common-utils": "^0.35.0",
     "@hyperdx/passport-local-mongoose": "^9.0.1",
     "@modelcontextprotocol/sdk": "^1.27.1",
     "@opentelemetry/api": "~1.9.1",

@@ -1,5 +1,26 @@
 # @hyperdx/app
 
+## 2.45.0
+
+### Minor Changes
+
+- b3cbd0260: feat: split the service map by up to 3 extra group-by SQL expressions (for
+  example `ResourceAttributes['k8s.pod.name']`), so a node is a service plus
+  those values and a slow or failing pod is no longer averaged into its service.
+
+### Patch Changes
+
+- ac913ceb3: feat: Choose how a PromQL source on ClickHouse is evaluated (the Prometheus HTTP API or the `prometheusQuery` table functions) instead of probing the server on every query. Without a choice the server is still probed, but the answer is now kept for 5 minutes. A source set to the HTTP API against a server that does not serve it fails with a clear error. Onboarding pins a PromQL source it detects to the HTTP API when the server serves it.
+- Updated dependencies [c17fc16a2]
+- Updated dependencies [1fba1c381]
+- Updated dependencies [7b03880eb]
+- Updated dependencies [2d17de5d7]
+- Updated dependencies [ac913ceb3]
+- Updated dependencies [7cfd333ea]
+- Updated dependencies [db4e7ba67]
+  - @hyperdx/api@2.45.0
+  - @hyperdx/common-utils@0.35.0
+
 ## 2.44.0
 
 ### Minor Changes

@@ -1,6 +1,6 @@
 {
   "name": "@hyperdx/app",
-  "version": "2.44.0",
+  "version": "2.45.0",
   "private": true,
   "license": "MIT",
   "engines": {
@@ -40,7 +40,7 @@
     "@dnd-kit/utilities": "^3.2.2",
     "@hookform/resolvers": "^3.9.0",
     "@hyperdx/api": "workspace:*",
-    "@hyperdx/common-utils": "^0.34.0",
+    "@hyperdx/common-utils": "^0.35.0",
     "@mantine/core": "^9.0.0",
     "@mantine/dates": "^9.0.0",
     "@mantine/dropzone": "^9.0.0",

@@ -41,7 +41,7 @@
   },
   "devDependencies": {
     "@clickhouse/client": "1.23.0-head.fae5998.1",
-    "@hyperdx/common-utils": "^0.34.0",
+    "@hyperdx/common-utils": "^0.35.0",
     "@jest/globals": "^30.2.0",
     "@types/crypto-js": "^4.2.2",
     "@types/jest": "^29.5.14",

@@ -1,5 +1,17 @@
 # @hyperdx/common-utils
 
+## 0.35.0
+
+### Minor Changes
+
+- 2d17de5d7: feat: serve PromQL label lookups from ClickHouse's Prometheus HTTP API
+
+  On ClickHouse 26.9+ with the `prometheus_api_v1` handler, `/labels` and `/label/:name/values` for a ClickHouse-backed PromQL source now proxy to the handler instead of querying the TimeSeries tags table with SQL. Older servers, servers without the handler, and sources set to `table-function` keep the SQL path, so a source that runs queries over HTTP on 26.6-26.8 still gets label autocomplete.
+
+### Patch Changes
+
+- ac913ceb3: feat: Choose how a PromQL source on ClickHouse is evaluated (the Prometheus HTTP API or the `prometheusQuery` table functions) instead of probing the server on every query. Without a choice the server is still probed, but the answer is now kept for 5 minutes. A source set to the HTTP API against a server that does not serve it fails with a clear error. Onboarding pins a PromQL source it detects to the HTTP API when the server serves it.
+
 ## 0.34.0
 
 ### Minor Changes

@@ -1,7 +1,7 @@
 {
   "name": "@hyperdx/common-utils",
   "description": "Common utilities for HyperDX application",
-  "version": "0.34.0",
+  "version": "0.35.0",
   "license": "MIT",
   "private": true,
   "files": [

@@ -1,5 +1,7 @@
 # @hyperdx/otel-collector
 
+## 2.45.0
+
 ## 2.44.0
 
 ### Patch Changes

@@ -1,7 +1,7 @@
 {
   "name": "@hyperdx/otel-collector",
   "description": "HyperDX OpenTelemetry Collector configuration and Docker image",
-  "version": "2.44.0",
+  "version": "2.45.0",
   "license": "MIT",
   "private": true
 }
@@ -4772,7 +4772,7 @@ __metadata:
     "@clickhouse/clickstack-node": "npm:^1.0.2"
     "@clickhouse/client-common": "npm:1.23.0-head.fae5998.1"
     "@esm2cjs/p-queue": "npm:^7.3.0"
-    "@hyperdx/common-utils": "npm:^0.34.0"
+    "@hyperdx/common-utils": "npm:^0.35.0"
     "@hyperdx/passport-local-mongoose": "npm:^9.0.1"
     "@modelcontextprotocol/sdk": "npm:^1.27.1"
     "@opentelemetry/api": "npm:~1.9.1"
@@ -4870,7 +4870,7 @@ __metadata:
     "@eslint/compat": "npm:^2.0.0"
     "@hookform/resolvers": "npm:^3.9.0"
     "@hyperdx/api": "workspace:*"
-    "@hyperdx/common-utils": "npm:^0.34.0"
+    "@hyperdx/common-utils": "npm:^0.35.0"
     "@mantine/core": "npm:^9.0.0"
     "@mantine/dates": "npm:^9.0.0"
     "@mantine/dropzone": "npm:^9.0.0"
@@ -4988,7 +4988,7 @@ __metadata:
   resolution: "@hyperdx/cli@workspace:packages/cli"
   dependencies:
     "@clickhouse/client": "npm:1.23.0-head.fae5998.1"
-    "@hyperdx/common-utils": "npm:^0.34.0"
+    "@hyperdx/common-utils": "npm:^0.35.0"
     "@jest/globals": "npm:^30.2.0"
     "@types/crypto-js": "npm:^4.2.2"
     "@types/jest": "npm:^29.5.14"
@@ -5015,7 +5015,7 @@ __metadata:
   languageName: unknown
   linkType: soft
 
-"@hyperdx/common-utils@npm:^0.34.0, @hyperdx/common-utils@workspace:packages/common-utils":
+"@hyperdx/common-utils@npm:^0.35.0, @hyperdx/common-utils@workspace:packages/common-utils":
   version: 0.0.0-use.local
   resolution: "@hyperdx/common-utils@workspace:packages/common-utils"
   dependencies: