sorratheorc/acs-logging-test

A set of simple tests designed to simluate long running logging processes along with periodic analysis of those logs. This code has been written to run as ongoing tests for the Azure Container Service but they may be useful in other test scenarios.

acs-logging-test's Issues

Build automated scaling

At present, the basic load test is manually scaled up. This results in the message queue being emptied and so multiple of the analyzers stop. This is good, we are scaling down well. However, we need to manually scale up again in order to stop the queue growing again.

We need to implement a monitor script that will test when the queue is growing and will start new analyzers accordingly.

Create some scripts to do load (basic) testing

We have the ability to start up one or more log producer and one or more log analyzers. We should build some scripts that will fire up a great many producers, watch the queue grow, then fire multiple analyzers to reduce the queue.

Use Azure File Service rather than mounted volumes

If deployed using Swarm all containers have to be scheduled on the same node as they share volumes. We should use Azure File Service instead.

Change Slack Channels to Env Variables

I think it would be a little easier to change Slack channels by migrating from hard coded values in notify.py to the environment variables. I can submit a pull, but I'd welcome any other feedback first.

Move config secrets to environment variables

Storage account keys need to be secret, but right now they are stored in the config.py file.

Add compose file

The demo scripts right now uses loops to scale up and down. It would be much better if we used docker-compose instead.

Use Azure Storage Queue

Log messages should be written to and read from Azure Storage Queue.

See http://justazure.com/microsoft-azure-storage-queues-part-1-getting-started/