cryptonomic / nautilus-cloud-metering Goto Github PK

Simplify the build process such that

• We do not require multiple shell scripts to containers
• Reduce the number of containers

The current design stores data at the per request/event level. This is too granular to query against.

Add a CQ to Influx such that events are aggregated to a suitable time window e.g. 1 minute or 5 minute intervals.

Reference : https://docs.influxdata.com/influxdb/v1.7/query_language/continuous_queries/
For playing around, use Time Series Admin for a GUI interface to InfluxDB.

Related to Cryptonomic/Nautilus-Cloud#74

Add travis CI integration to this project

The following topics need some love in the documentation:

• Test and update docker artillery instructions
• Add a blurb on updating the proxy nginx configuration in docker

Currently metering API returns all aggregated data for continuous queries.
We want to limit the amount of records fetched from given timestamp. For example:

/queries/5m?from=123456789

Nautilus Cloud should be able to meter requests to the various services it provides. To provide this functionality create a service with the following constraints:

1. Be able to meter any application or service without necessitating modification to the application or service being metered.
2. Enforce metering policy as soon as the request hits the Nautilus cloud ecosystem.
3. Be able to provide usage statistics to the end user.
4. Be lightweight and flexible.

The APIKey authorization logic should allow options request to pass through.

Static keys stop working when keys are fetched from Nautilus Cloud.

Configuration for Nautilus-Cloud-Metering Agent should be read from an external file. Currently, the configuration needs to be put into src/main/resources/ folder and is embedded into the JAR file after building.

This is bad because:

• any change in the config file will require rebuilding JAR file
• secrets will be embedded into the JAR file in plain text form