LGTM. We can remove the `google_` prefix of the configuration as the plugin name already explained it.
JinChao Shuai <[email protected]> 于2021年11月11日周四 上午11:42写道： > > ---- > Configuration： > > { > "google_auth_config": { > "private_key": "***" > "client_email": "***" > "project_id": "***" > "token_uri": "***" > }, > "google_auth_file": "/path/to/google-service-account.json", > "resource": { > "type": "global" > }, > "log_id": "cloudresourcemanager.googleapis.com%2Factivity", > "batch_max_size": 200, > "batch_max_size": 10, > } > > // google_auth_config // the google service account > config（Semi-optional, one of `google_auth_config` or `google_auth_file` > must be configured） > // google_auth_config.private_key // the private key parameters of the > Google service account > // google_auth_config.client_email // the client email parameters of the > Google service account > // google_auth_config.project_id // the project id parameters of the > Google service account > // google_auth_config.token_uri // the token uri parameters of the > Google service account > // google_auth_file // path to the google service account > json file（Semi-optional, one of `google_auth_config` or `google_auth_file` > must be configured） > // resource // the Google monitor resource, refer > to: > https://cloud.google.com/logging/docs/reference/v2/rest/v2/MonitoredResource > // log_id // google logging id, refer to: > https://cloud.google.com/logging/docs/reference/v2/rest/v2/LogEntry > // batch_max_size // batch queue max number of bearer > entries > // batch_timeout // batch queue timeout of bearer entries > > > ---- > Details: > > 1. Obtain and assemble request information in the APISIX Log stage > 2. To interact with google logging service for the first time, you need to > request token information. After obtaining the token, it will be cached in > the memory of the working node. > 3. After obtaining a valid token, put the request information into the > batch processing queue. When the batch processing queue triggers the > batch_max_size or batch_timeout threshold, the data in the queue is > synchronized to the google cloud logging service > 4. Before each request is sent, check whether the token is about to time > out, and refresh the token if it will time out. > > Zhiyuan Ju <[email protected]> 于2021年11月11日周四 上午10:01写道： > > > Yes, kindly let us know when the proposal details are updated :) > > > > Best Regards! > > @ Zhiyuan Ju <https://github.com/juzhiyuan> > > > > > > Zexuan Luo <[email protected]> 于2021年11月10日周三 下午6:30写道： > > > > > Yes. But where are the configuration and technology details of the > > > feature? It seems like a feature request instead of a proposal. > > > > > > JinChao Shuai <[email protected]> 于2021年11月10日周三 下午5:03写道： > > > > > > > > Hi, community. > > > > > > > > Currently, the cloud service provider that APISIX stores analysis logs > > > only > > > > supports Alibaba Cloud. > > > > > > > > As one of the world's largest cloud service providers, Google has a > > very > > > > large user base. > > > > > > > > Therefore, I propose that APISIX supports the synchronization of logs > > to > > > > Google Cloud Logging in the form of plug-ins, which can not only meet > > the > > > > diverse log storage and analysis needs of users, but also enrich the > > > > surrounding ecology of APISIX. > > > > > > > > > > > > > > > > This is my suggestion [1], developers are welcome > > > > Join the discussion about its suggestions and questions. > > > > > > > > [1] https://github.com/apache/apisix/issues/5474 > > > > > > > > -- > > > > Thanks, > > > > Janko > > > > > > > > -- > Thanks, > Janko
