Tencent Cloud CLS incident alerts
Availability: preview
Tencent Cloud CLS uses the same public webhook endpoint as generic alerts, with provider-specific validation and evidence handling. This guide covers the implemented callback path; it does not claim a general CLS polling connector.
Configure the trigger
Section titled “Configure the trigger”After signing in, save a signed_webhook trigger and choose provider
tencent_cls in the versioned trigger configuration. Configure
MENDRY_PUBLIC_URL, then copy the generated inbound URL into the CLS alarm
callback destination. Keep the URL token private.
The callback must use a supported media type and valid Tencent callback shape. Unknown, disabled, incomplete, or malformed token configurations return the same not-found boundary.
Ingestion and grouping
Section titled “Ingestion and grouping”Mendry validates the callback envelope before returning 202. Tencent grouping
uses a deterministic fingerprint from validated provider fields; model output,
account metadata, topic IDs, and callback URLs do not become grouping authority.
The stored Observation retains the accepted body, while normalized alert
evidence stores a bounded projection and provenance.
Trusted detail evidence
Section titled “Trusted detail evidence”A remediation run can follow an approved Tencent detail URL only through the trusted adapter. The adapter validates the regional host and response, strips callback/secret-bearing fields, and persists bounded evidence. A Tencent-backed run cannot claim successful diagnosis until required direct provider detail is available and cited.
Failure handling
Section titled “Failure handling”Inspect safe webhook.ingest.failed and Tencent request/detail diagnostics.
Logs omit inbound tokens, callback URLs, H5 shield fields, secret IDs, secret
text, and malformed control content.
Verified against
Section titled “Verified against”backend/internal/modules/hooks/adapter/tencentcls/client.go