Wednesday, October 15, 2014

Listen to The Cloudcast

The Cloudcast #165 - DevOps Automation as a Service






Topic 1 - Tonight is interesting because we are pulling in pieces from a bunch of previous podcasts and past guest topics. We have spoken about the many emerging trends and use cases in DevOps, but a big problem has been how do you put all the pieces together. You hear about Jenkins, Docker, OpenStack, PagerDuty, Loggly, AWS, etc. A lot of moving pieces that we have to integrate together and then actually operate efficiently.


Let’s start at the start, what is the concept of Automation as a Service and what problem are we trying to solve?


Topic 2 - Quote from Blog: "because developers are in charge that every single API must be a first class citizen. They determine whether your API is inadequate very quickly. If you treat your APIs badly by deprecating them suddenly and without warning, you are essentially slapping developers that use your APIs in the face." - Very true. There is a presentation your company did on OpenStack vs. VMware and the idea of closed vs. open with some great analogies to history and advancements in efficient. How does AaaS help developers?


Topic 3 - Third Wave in IT. You present a potential framework:
  1. Docker automating tracking of all dependencies for an environment while providing efficient and very fast to deploy containers; 
  2. Jenkins automating QA testing; 
  3. OpenStack and Docker orienting orchestration solutions 
  4. Monitoring 
  5. StackStorm and others are focusing on automation as a service - Remediation through automation?
Topic 4 - Is AaaS the “glue” between a bunch of existing projects and frameworks to create an automation workflow through change management, remote execution? Not trying to replace Docker, Salt, Ansible, Chef, OpenStack, Jenkins, New Relic? Aren’t all these integrations points a nightmare? Is this an on-prem product, cloud offering?

Topic 5 - You mention machine learning and artificial intelligence. This sounds a bit like what VMturbo tries to do at the hypervisor level. In discussion with them I know getting people flipping the bit that enables full automation makes some folks uncomfortable. They start to think of SkyNet in Terminator and the machines taking over the world. Thoughts?

Thursday, September 25, 2014

The Cloudcast #162 - Building and Managing Scalable SaaS Services




Topic 1 - Manoj, you and your team came highly recommended to us by the team at Evident.io (Tim Prendergast) and we learned about your service at AWS Summit in NYC. Tell us about your background and how it eventually led you to Loggly.


Topic 2 - You have an excellent talk/presentation on Critical SaaS Mistakes to Avoid. You mention that scalability needs to be priority #1. How much different is building applications/service in the cloud vs. building packaged software?


Topic 3 - We presume that Loggly was built from Day 1 was a web-scale SaaS application. Having built it, what might you do differently or major lessons learned? Realistically, is it possible for someone to SaaS-ify an existing application?


Topic 4 - Let’s talk about Loggly. Every company, every application has logs and they are a cluttered mess of potentially valuable information. People throw them at Loggly. What happens next?


Topic 5 - That has to be a really complex system on the backend to be able to ingest, parse, analyze, tag all the data - keep it isolated by customer - manage historical logs - then visualize it and give recommendations in real-times. Can you give us some sense of what goes on behind the scenes?


Topic 6 - Logging became somewhat more visible at AWS Summit when AWS announced centralized log management. How does your world change when AWS elevates a service that is in your domain?

Topic 7- What are the most common scenarios where companies decide they need help with log management?