The code consist of 3 components:

• db - contains a docker compose setup for running a MongoDB instance, that will stores the timeseries data
• ingestion - a script that is run over the DB (ingestion is run from within the docker container, after the file is copied over)
• api - an api that allows querying the data for specific time ranges and aggregations by frequency period

Each module describes the running instructions needed, but in order for you to run the code, you need:

• Golang installed (>1.7)
• Docker installed
• bash shell

1. Start the docker container from within the db directory with:
   ./run.sh
   After you finished, stop the container with:
   ./stop.sh

2. Run the ingestion script - any time you consider it. In an ideal future there would be a cronjob set up that runs it every 5 minutes to observe/save behaviour continously. The ingestion script should also be run from the ingestion folder: ./ingest.sh

3. Start the API. The Api can be started even before the ingestion layer has run.
   go run main.go

Some interesting queries that you can run:

curl "localhost:8080/metrics/concurrency?start=1501681460&end=1650843741&frequency=years" | jq
Returns all the saved concurency metrics within the given time range, with aggregations done /years, displaying the avarage for each. Results could look like:

[
  {
    "timestamp": "2017-01-01T00:00:00Z",
    "concurrency": 169070
  },
  {
    "timestamp": "2022-01-01T00:00:00Z",
    "concurrency": 263415
  }
]

curl "localhost:8080/metrics/cpu_load?start=1501681460&end=1650843741&frequency=minutes"
curl "localhost:8080/metrics?start=1501681460&end=1650843741&frequency=minutes"

curl "localhost:8080/metrics/average?start=1501681460&end=1650843741"
curl "localhost:8080/metrics/cpu_load/average?start=1501681460&end=1650843741"
curl "localhost:8080/metrics/concurrency/average?start=1501681460&end=1650843741"

DB level:

• Authentication to Mongo should use secrets
• Authentication for the script executer needs to be set up

API level TODOs:

• change the Logging - use Logger
• revise parameters - http server, mongo client
• extend query methods with more aggregation; query for a day (without using the range)
• Mongo - data from Mongo; read into a channel and considering streaming it to the client or offer pagination
• write more tests (cover more testcases, and add tests for the handler repo)
• add a health endpoint

Ingestion Level:

• create a cron job
• make ingestion through a client, instead of copying the script inside the container and running it there