outline-server/src/metrics_server
Jonathan Cohen cf355a6fee
Per-Key Data Limits in Shadowbox (#769)
* s/AccessKeyDataLimit/DefaultDataLimit/g for shadowbox

* Implement per-key custom data limits in shadowbox

* Shadowbox integration test

* Fix the manager service test to maintain backwards compatibility with
the GET /server response

* Fix integration test for Travis

* Fix lint errors

* Implement removal of access key data limits

* Persist custom data limits

* Bump Shadowbox version number

* Add a per key dat alimit count to the feature metrics endpoint

This also involved modifying the schema of
uproxysite:uproxy_metrics_dev.feature_metrics to include the new field.
See https://cloud.google.com/bigquery/docs/managing-table-schemas#bq_1
for how this was accomplished

Tested:
  yarn do metrics_server/test && yarn do metrics_server/test_integration

* Collect usage metrics for per-key data limits

* API documentation and cleanup

* Respond to review comments

* Respond to review comments

* Don't encode the data limit in the access key url

This should have been caught in the integration test, but the
integration test was miswritten with "echo" instead of "fail", cause the
test to mistakenly pass.

Fixing this exposed other errors in the test, which were fixed.  The
test has been confirmed to both pass and fail successfully.

* Fix the manager service unit test

* Remove stray files

* Split integration test into individual functions

* Clarify comment in the AccessKey interface

* Be more explicit about behavior when no data limits are set

* Remove spuriosu import added by vscode

* DOn't bump shadowbox version number yet

* Rename to accessKeyToApiJson

* Don't export AccessKeyJson and use "StorageJson"

* Don't await data limit enforcement

* Rename the defaultDataLimit getter

* Make dataLimit in the accesskey model readonly

* Use nullish coalescing operator
2020-12-23 13:52:50 -05:00
..
infrastructure Feature metrics (#577) 2020-02-26 12:49:52 -05:00
app_dev.yaml Feature metrics (#577) 2020-02-26 12:49:52 -05:00
app_prod.yaml Feature metrics (#577) 2020-02-26 12:49:52 -05:00
build_action.sh make metrics server build script an action, update docs 2018-12-10 14:22:47 -05:00
config_dev.json Feature metrics (#577) 2020-02-26 12:49:52 -05:00
config_prod.json Feature metrics (#577) 2020-02-26 12:49:52 -05:00
connection_metrics.spec.ts Feature metrics (#577) 2020-02-26 12:49:52 -05:00
connection_metrics.ts Feature metrics (#577) 2020-02-26 12:49:52 -05:00
deploy_dev_action.sh Feature metrics (#577) 2020-02-26 12:49:52 -05:00
deploy_prod_action.sh Feature metrics (#577) 2020-02-26 12:49:52 -05:00
dispatch.yaml Feature metrics (#577) 2020-02-26 12:49:52 -05:00
feature_metrics.spec.ts Per-Key Data Limits in Shadowbox (#769) 2020-12-23 13:52:50 -05:00
feature_metrics.ts Per-Key Data Limits in Shadowbox (#769) 2020-12-23 13:52:50 -05:00
index.ts Update @types/express (#602) 2020-03-09 16:06:23 -04:00
model.ts Per-Key Data Limits in Shadowbox (#769) 2020-12-23 13:52:50 -05:00
package.json Select newly added keys (#601) 2020-03-06 18:33:07 -05:00
README.md Per-Key Data Limits in Shadowbox (#769) 2020-12-23 13:52:50 -05:00
run_action.sh Feature metrics (#577) 2020-02-26 12:49:52 -05:00
test_action.sh Fix unit tests (#693) 2020-06-22 16:32:48 -04:00
test_integration_action.sh Per-Key Data Limits in Shadowbox (#769) 2020-12-23 13:52:50 -05:00
tsconfig.json make metrics server build script an action, update docs 2018-12-10 14:22:47 -05:00

Outline Metrics Server

The Outline Metrics Server is a Google App Engine project that writes feature and connections metrics to BigQuery, as reported by opted-in Outline servers.

API

Endpoints

The metrics server deploys two services: dev, used for development testing and debugging; and prod, used for production metrics. The dev environment is deployed to https://dev.metrics.getoutline.org; the prod environment is deployed to https://prod.metrics.getoutline.org. Each environment posts metrics to its own BigQuery dataset (see config_[dev|prod].json).

URLs

The metrics server supports two URL paths:

  • POST /connections: report server data usage broken down by user.

    {
      serverId: string,
      startUtcMs: number,
      endUtcMs: number,
      userReports: [{
          userId: string,
          countries: string[],
          bytesTransferred: number,
      }]
    }
    
  • POST /features: report feature usage.

    {
        serverId: string,
        serverVersion: string,
        timestampUtcMs: number,
        dataLimit: {
            enabled: boolean
            perKeyLimitCount: number
        }
    }
    

Requirements

Build

yarn do metrics_server/build

Run

Run a local development metrics server:

yarn do metrics_server/run

Deploy

  • Authenticate with gcloud:
    gcloud auth login
    
  • To deploy to dev:
    yarn do metrics_server/deploy_dev
    
  • To deploy to prod:
    yarn do metrics_server/deploy_prod
    

Test

  • Unit test
    yarn do metrics_server/test
    
  • Integration test
    yarn do metrics_server/test_integration