Health check api by hazel-bohon · Pull Request #5952 · Particular/ServiceControl

hazel-bohon · 2026-10-05T16:13:12Z

Addresses the first part of Replace internal platform health custom checks with an explicit API #5860
Defines a new API endpoint for the frontend to use.

Copilot

Copilot review overview

🟡 Changes recommended

Five-second polling currently forces license discovery and persistence access for every open ServicePulse client.

Review effort: Balanced
Findings: 1

Open (1)

Avoid full license refresh on every platform-health poll · New

What changed in this PR

Adds a dedicated Platform Health API for ServicePulse, separating internal health reporting from customer custom checks.

Changes:

Adds platform health inventory, alerts, licensing, and remote-instance metadata.
Extends configuration discovery and remote probing.
Adds unit, acceptance, authorization, and API approval coverage.

Reviewed against issue #5860, docs/platform-health.md, and the linked ServicePulse models/store. No private context was provided.

File	Description
`src/ServiceControl/ServiceControlApiHostBuilderExtensions.cs`	Registers license information provider.
`src/ServiceControl/PlatformHealth/PlatformHealthState.cs`	Tracks internal health-check state.
`src/ServiceControl/PlatformHealth/PlatformHealthController.cs`	Exposes the health endpoint.
`src/ServiceControl/PlatformHealth/PlatformHealthApi.cs`	Builds platform health responses.
`src/ServiceControl/PlatformHealth.http`	Adds manual API requests.
`src/ServiceControl/Licensing/LicenseInfoProvider.cs`	Extracts shared license mapping.
`src/ServiceControl/Licensing/LicenseController.cs`	Uses the shared provider.
`src/ServiceControl/Infrastructure/WebApi/RemoteInstanceServiceCollectionExtensions.cs`	Preserves remote URL prefixes.
`src/ServiceControl/Infrastructure/Api/ConfigurationApi.cs`	Adds discovery metadata and robust remote probing.
`src/ServiceControl/CustomChecks/CustomChecksComponent.cs`	Registers platform health services.
`src/ServiceControl/CustomChecks/CustomCheckResultProcessor.cs`	Records internal health reports.
`src/ServiceControl.UnitTests/ScatterGather/RemoteInstanceHttpClientTests.cs`	Tests remote configuration requests.
`src/ServiceControl.UnitTests/PlatformHealth/PlatformHealthStateTests.cs`	Tests health-state behavior.
`src/ServiceControl.UnitTests/PlatformHealth/PlatformHealthApiTests.cs`	Tests response assembly and failures.
`src/ServiceControl.UnitTests/Licensing/ActiveLicenseTests.cs`	Tests extracted license mapping.
`src/ServiceControl.UnitTests/ApprovalFiles/APIApprovals.RootPathValue.approved.txt`	Approves discovery URL addition.
`src/ServiceControl.UnitTests/ApprovalFiles/APIApprovals.HttpApiRoutes.approved.txt`	Approves the new route.
`src/ServiceControl.UnitTests/API/APIApprovals.cs`	Tests primary configuration metadata.
`src/ServiceControl.MultiInstance.AcceptanceTests/Infrastructure/When_inspecting_platform_health.cs`	Exercises multi-instance health behavior.
`src/ServiceControl.Audit/Infrastructure/WebApi/RootController.cs`	Publishes audit identity metadata.
`src/ServiceControl.Audit.UnitTests/API/APIApprovals.cs`	Tests audit configuration metadata.
`src/ServiceControl.Api/IPlatformHealthApi.cs`	Defines the health API abstraction.
`src/ServiceControl.Api/Contracts/RootUrls.cs`	Advertises platform health discovery.
`src/ServiceControl.Api/Contracts/PlatformHealthView.cs`	Defines response contracts.
`src/ServiceControl.AcceptanceTests/WebApi/When_the_configuration_page_is_read.cs`	Verifies configuration metadata.
`src/ServiceControl.AcceptanceTests/Security/OpenIdConnect/When_authentication_is_enabled.cs`	Verifies endpoint authorization.
`src/ServiceControl.AcceptanceTests/Monitoring/CustomChecks/When_custom_checks_are_classified.cs`	Verifies health/custom-check separation.
`docs/README.md`	Indexes the design documentation.
`docs/platform-health.md`	Documents the API contract and rationale.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

+        var hasConnector = connectorHeartbeatStatus.LastHeartbeat != null;
+        try
+        {
+            var license = await licenseInfoProvider.GetLicense(true, "servicepulse", cancellationToken);


Co-authored-by: hazel-bohon <2416062+hazel-bohon@users.noreply.github.com>

mchamberlin77 · 2026-10-06T17:32:20Z

+
+Association uses case-insensitive instance name plus reporting host ID. A legacy remote without a host ID can use a name match only when there is one matching inventory row and one reporting host with that name. Ambiguous or unmatched reports remain in root `alerts` without `instance_id`; they are never assigned to several rows. Consumers should retain a place to display those unassigned alerts.
+
+The legacy summary describes captured checks, not the whole browser-visible platform: `status` is `unknown` before any internal report, `healthy` when none are failing, and `unhealthy` when at least one is failing. Its corresponding `severity` values are `unknown`, `none`, and `error`. ServicePulse should use per-instance health for page severity and combine it with its independently observed monitoring state. The legacy summary does not account for monitoring, browser connectivity, license expiry, or available upgrades.


This seems to conflict with the health field above that gives degraded for instances with failures.

How is severity determined? I do not see anything in the table above that gives an indication of issue serverity.

mchamberlin77 · 2026-10-06T17:34:19Z

+| `transport_type`, `error_queue`, `error_log_queue`, `forward_error_messages` | Available transport configuration; a known `false` forwarding setting is preserved |
+| `audit_queue`, `audit_log_queue`, `forward_audit_messages` | Available audit transport configuration |


Are these multiple fields listed in the same row? They read like the possible values for a field and not the field name themselves.

mchamberlin77 · 2026-10-06T17:39:44Z

+
+### License
+
+`license.availability` is `available` after a successful refresh and `unavailable` when license details cannot be refreshed. An unavailable license never claims to be valid and does not suppress instance health.


Is the license something currently checked by the Custom Checks health checks? From what I see in the task adding a license api is out of scope.

If we want to expand scope and add this it could be in a separate PR to keep the work for each new API clear.

mchamberlin77 · 2026-10-06T17:47:43Z

 - [Retries over Azure Storage Queues transport](retries-asq-transport.md) — transport-specific retry handling
 - [Data versioning design](data-versioning-design.md) — the cache-versioning invariant for API responses
 - [Event log design](eventlog-design.md) — what the event log is and what it records
+- [Platform health API](platform-health.md) — how ServicePulse reads internal health independently from customer custom checks


Suggested change

- [Platform health API](platform-health.md) — how ServicePulse reads internal health independently from customer custom checks

- [Platform health API](platform-health.md) — how ServicePulse can read internal health independently from customer custom checks

mchamberlin77 · 2026-10-06T17:53:43Z

+        public Guid HostId { get; set; }
+    }
+
+#nullable enable


Why is this not applied for the entire file?

hazel-bohon added 2 commits September 30, 2026 15:04

first pass

91735c5

Separate internal checks from the custom checks

8ecd9bb

internalautomation Bot assigned hazel-bohon Oct 5, 2026

This comment was marked as duplicate.

Sign in to view

hazel-bohon requested review from afprtclr and mchamberlin77 and a balanced review from Copilot October 5, 2026 16:15

Copilot started reviewing on behalf of hazel-bohon October 5, 2026 16:16 View session

Copilot AI reviewed Oct 5, 2026

View reviewed changes

Comment thread src/ServiceControl/PlatformHealth/PlatformHealthApi.cs

var hasConnector = connectorHeartbeatStatus.LastHeartbeat != null;

try

{

var license = await licenseInfoProvider.GetLicense(true, "servicepulse", cancellationToken);

Copilot started work on behalf of hazel-bohon October 6, 2026 11:51 View session

Pin Azure CLI on Windows CI

e1ceaa9

Co-authored-by: hazel-bohon <2416062+hazel-bohon@users.noreply.github.com>

Copilot finished work on behalf of hazel-bohon October 6, 2026 12:00

mchamberlin77 reviewed Oct 6, 2026

View reviewed changes

Navigation Menu

Health check api - #5952

Health check api#5952
hazel-bohon wants to merge 3 commits into
masterfrom
health-check-api

hazel-bohon commented Oct 5, 2026 •

edited

Loading

Uh oh!

This comment was marked as duplicate.

Uh oh!

Copilot AI left a comment

Uh oh!

mchamberlin77 Oct 6, 2026

Uh oh!

mchamberlin77 Oct 6, 2026

Uh oh!

mchamberlin77 Oct 6, 2026

Uh oh!

mchamberlin77 Oct 6, 2026

Uh oh!

mchamberlin77 Oct 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants


		Association uses case-insensitive instance name plus reporting host ID. A legacy remote without a host ID can use a name match only when there is one matching inventory row and one reporting host with that name. Ambiguous or unmatched reports remain in root `alerts` without `instance_id`; they are never assigned to several rows. Consumers should retain a place to display those unassigned alerts.

		The legacy summary describes captured checks, not the whole browser-visible platform: `status` is `unknown` before any internal report, `healthy` when none are failing, and `unhealthy` when at least one is failing. Its corresponding `severity` values are `unknown`, `none`, and `error`. ServicePulse should use per-instance health for page severity and combine it with its independently observed monitoring state. The legacy summary does not account for monitoring, browser connectivity, license expiry, or available upgrades.

		\| `transport_type`, `error_queue`, `error_log_queue`, `forward_error_messages` \| Available transport configuration; a known `false` forwarding setting is preserved \|
		\| `audit_queue`, `audit_log_queue`, `forward_audit_messages` \| Available audit transport configuration \|


		### License

		`license.availability` is `available` after a successful refresh and `unavailable` when license details cannot be refreshed. An unavailable license never claims to be valid and does not suppress instance health.

	- [Platform health API](platform-health.md) — how ServicePulse reads internal health independently from customer custom checks
	- [Platform health API](platform-health.md) — how ServicePulse can read internal health independently from customer custom checks

Conversation

hazel-bohon commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment was marked as duplicate.

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

mchamberlin77 Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

mchamberlin77 Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

mchamberlin77 Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

mchamberlin77 Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

mchamberlin77 Oct 6, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

hazel-bohon commented Oct 5, 2026 •

edited

Loading