Home of quality reports and checks that are run daily at 2AM.

A good report will be responsible for doing a handful of things:

• Use defined metrics and clear visualizations, when applicable
• Alerting stakeholders of interesting or erroneous data
• Save the numbers and figures in easy to parse and retrieve formats

Quantities should be defined cleary, (Users vs Number of Users), and consistently (all rows have same dimension, no nested tables). Visualization should follow good practices. See this (guide or this one for more.

A report may return attachments that will be sent to slack after the report has been compiled. This is useful for alerting users of interesting data or bring attention to problems.

Each report is given an output path in s3 where it is expected to dump its results. A report may also return a list of local files to be uploaded directly to slack, making them even more accessible.

Reports are added as new modules within the reports module. They are expected to have a handler(event, context) function which acts similar to the standard aws lambda handler. To add a new report to the daily run, add a new entry in the reports inside the invoker.py handler() function with the name and module of the new report

The handler(event, context) function of the report is expected to return a tuple of (attachments, files). The attachments is a list of Slack attachments that will be sent to slack after the report is run and the files is a dict of {name: path} entries where the file in each path in the report's environment will be uploaded to slack and titled with name as well as uploaded to s3.

Here is a basic example report function demonstrating the minimum format:

def handler(event, context):
    attachments = [ { "text": "Hello World" } ]

    files = {
        'Many digits of pi': '/tmp/pi.txt'
    }

    return attachments, files