Single service comparison

Cloud Logging models data as a single global log store per project/folder with log-based metrics, exclusion filters, and a Logging router that fans entries out to Cloud Storage, Pub/Sub, and BigQuery. Huawei LTS organizes data into log groups and log streams, ingests 40+ native Huawei services (ELB, VPC, CTS, APIG, WAF) plus API/SDK/cross-cloud/device logs, and transfers to OBS, DMS, and DWS; it has no direct equivalent of project-level log-based metrics or exclusion filters, so metric derivation and volume control must be rebuilt with AOM/CES and stream-level ingestion rules.

GCP exposes a unified REST/gRPC Logging API with consistent entry fields (logName, severity, jsonPayload, resource, labels) and tight IAM/Bucket ACL control over sinks. LTS uses its own LTS APIs plus ICAgent for host collection; schema mapping and labeling must be translated from GCP's monitored-resource types to LTS log streams and structured-parsing templates, and IAM semantics differ—LTS access is governed by Huawei IAM on log groups/streams rather than GCP bucket-level sink ACLs.

Operationally, Cloud Logging offers serverless ingestion at planet scale, configurable retention/custom buckets, and log-based alerting integrated with Cloud Monitoring. LTS scales horizontally to hundreds of TB/day with tiered hot/cold storage and SLA of 99.95%, alerts on keywords or SQL via SMS/email/DingTalk/WeCom/HTTP, and relies on CTS for audit-grade trails rather than GCP's unified Cloud Audit Logs integration. Integration depth, alert expression language, and audit pipeline therefore differ materially and must be re-planned rather than re-pointed.

Assess current usage: enumerate GCP log sinks, monitored-resource types, retention and custom buckets, exclusion filters, log-based metrics, and alert policies; map each to LTS log groups/streams, structured-parsing templates, SQL/keyword alarm rules, and AOM/CES for metric and event parity. Confirm regional availability of LTS in your target Huawei region and that ICAgent supports your host OS and runtime before committing, as LTS is the core equivalent but not a like-for-like replacement for the GCP router and metrics model.

For data and application migration, replace the GCP Logging agent/API calls with ICAgent on VMs/containers and the LTS ingestion API/SDK for custom sources; translate jsonPayload and labels into LTS structured fields and parsing templates. Re-implement sink fan-out using LTS transfer tasks to OBS (long-term retention, mirroring Cloud Storage sinks), DMS (for Pub/Sub-like streaming), and DWS (for BigQuery-like analytics), and rebuild log-based metrics as AOM/CES custom metrics since LTS does not emit derived metrics directly.

Validate parity before cutover: run GCP and Huawei ingestion in parallel for a representative window, compare completeness, latency, search-export row counts, and alert-trigger behavior, and confirm audit coverage with CTS trailing GCP's Cloud Audit Logs. Recompute TCO using Huawei LTS ingestion, indexing, read/write, tiered storage, and transfer volume plus AOM/APM/CES where used, instead of GCP ingestion/retention/query pricing, sizing for peak log rate and retention period; cut over only after alert and dashboard equivalents pass in the Huawei console and downstream consumers point at DMS/OBS/DWS.