CloudShell monitorability

We need best practices or a mechanism for automated monitoring of the CloudShell infrastructure. While the CloudShell Monitor is a good tool for looking at logs and alerts interactively, there is no way to feed anything from it into a proper monitoring tool that can generate alerts, keep a history, etc.

I think there are a few ways this can go. One is to make a CloudShell Monitor service, with a REST API. Other monitoring systems could query the API for events and metrics.

Another option is for Quali to publish best practices on monitoring the CS infrastructure. These would include:

• Which logs and log directories to monitor
  • Which lines indicate problems, versus which are more informational
• Which processes should be running
• Which ports should be open
  • Test queries for interacting with some processes
• Hooks for obtaining performance metrics - e.g. how long environment and resource driver commands are taking

Data Dog has a good summary of effective monitoring.

There is a third option, which I only mention here because I don't like it and want to explain why, and that is for Quali to provide a closed, preconfigured monitoring system with the product. I think this would be difficult to implement in a way that accounts for all the deployable architectures, and it would lock both Quali and its customers into a single platform. Also, doing this instead of exposing data and logs goes against the openness that makes CloudShell such a powerful solution.