In order to properly observe and track swatch-producer-azure we must log state changes and errors more thoroughly. When error cases occur we should alert on them from splunk.

Important fields to log: 

• org_id
• billing_account_id
• billing_provider
• billing_factor
• value
• product_id
• metric_id
• remittance_uuid
• aggregate_id
• remittance timestamp
• aggregate window timestamp

Fields from aws:

• UsageEventId
• Dimension
• Value
• Timestamp

Areas to log:

 Normal flow:

• Consumed billable-usage-aggregate
• Reason for not sending to marketplace if not sent(ex: metric_id invalid)
• Request to azure api
• Response from azure api
• Status changes sent to billable-usage.status topic

Error cases:

• Subscription missing
• Api failures from swatch-contracts
• Api failures from azure marketplace

Alerts: 

• Customer not subscribed when no contract found
• Failed message rate increase (5 in 5 minutes?)