Which cloud-native technologies appear in the Ekuaibao story?

The published story names Kubernetes, Istio, Knative, PaaS, and the previous ELK, Prometheus, and third-party APM tools. Components and versions depend on the customer environment.

How do tags connect users to infrastructure?

Consistent user, service, and environment tags can link request traces with logs, hosts, Pods, and processes, provided the collection and field conventions remain aligned.

Does a unified platform require removing every open-source component?

Not necessarily. The published story describes consolidation and replacement of separate monitoring entry points. Migration scope should follow dependency, retention, and risk assessment.

Ekuaibao: Kubernetes, Istio & End-to-End Observability

Customer context

Ekuaibao is an enterprise spend, expense-control, and reimbursement platform. Its published story describes microservices running on Kubernetes with technologies including Istio and Knative, closely coupled to PaaS and business systems.

The cloud-native stack added dynamic relationships

Kubernetes, Istio, Knative, and PaaS created changing service, Pod, process, and network relationships that required consistent tags.

ELK, Prometheus, and APM were separate

Logs, metrics, and application traces could not retain user and request context across tools, adding breaks to both maintenance and investigation.

Implementation

Unify collection and analysis with DataKit

DataKit brought telemetry from different stacks into one platform so collection, queries, visualisation, and alerts could share object tags.

Connect the user path with tags

User attributes and trace data were linked through tags, then correlated to Service, Log, host, Pod, and process data.

Consolidate separate open-source monitors

A shared platform took over separate monitoring entry points, reducing context switching and repeated maintenance across systems.

What changed

Multiple stacks entered one operating view

Teams could inspect related metrics, logs, traces, and runtime objects by service, user, environment, and time range.

Roles shared investigation context

Developers, platform engineers, and operators could collaborate around the same request and object relationships.

The monitoring maintenance surface narrowed

After consolidating duplicate collection, query, and display entry points, teams could focus more effort on data quality and reliability workflows.

Ekuaibao Guance