Google’s Evolving Cloud Ecosystem and the Challenge for DevOps
Last week, Google announced several changes to its cloud platform. First, AppScale, the company that provides an open source implementation of Google’s application platform, Google App Engine, is receiving a direct investment from Google in order to accelerate the interoperability between AppScale and Google App Engine. This is a smart move, and it should help developers overcome the app portability issue that is ushering in a new era of vendor-lock within public clouds.
Google App Engine is a Platform-as-a-Service (PaaS) offering for building and running applications on Google’s infrastructure. App Engine applications are intended to be easy to build, easy to maintain, and easy to scale as app traffic and data storage needs change.
With AppScale, developers can develop web and mobile applications using App Engine APIs and run the apps on any cloud – public or private.
In a blog post on Tuesday, Miles Ward, Global Head, Solutions for Google Cloud Platform, said, “Imagine you have a subset of customers who would be better served by your infrastructure because they have custom integration requirements. You can continue to serve your worldwide customers with the power of App Engine, and route requests from specific customers down to an installation of AppScale just for them. Imagine you have an existing data center or colocation capacity, but want to get started building apps designed with the cloud in mind. Simply install AppScale today and start building apps that are cloud-ready.”
What DevOps Teams Need to Know
We at BigPanda believe in and embrace open-source-based innovations. The ability to deploy apps wherever and whenever they best serve users is undeniably beneficial.
However, DevOps teams need to be ready to manage the sprawl. As apps creep beyond their traditional boundaries behind corporate firewalls into new and hybrid infrastructures, DevOps teams need to figure out how to monitor, maintain, and manage these disparate systems.
Google made another announcement last week that, at first glance, seems to address these concerns.
“Having trouble tracking down a bug in production when time is of the essence? Cloud Debugger can help you find the culprit in a few clicks. With no configuration required, simply select a line of code and let Cloud Debugger return the local variables and a full stack trace when that line is next executed – all without halting your application or slowing down any requests,” wrote Keith Smith, Product Manager at Google.
That’s great, but there’s a step missing here. Before you can troubleshoot your code, you need to know that there’s a problem in the first place, and that’s where DevOps teams struggle.
The write-once, run-anywhere ethos that AppScale (and now Google to some degree) aspires to is a good thing, but there needs to be a similar move to monitor-once, remediate-everywhere with IT incident response tools. Otherwise, the flood of alerts coming in from AWS, Nagios, New Relic, Splunk, and so forth will paralyze DevOps teams.
As apps migrate beyond their traditional boundaries, and as they now migrate beyond even their new cloud boundaries, service assurance becomes a moving target. If an app has an issue in AWS, will the problem be mirrored on your on-premises deployment? It’s hard to say, and without a unified, centralized approach to IT incident management, you’ll be forced to make educated guesses, at best.
Agile Incident Sharing
If you’ve read this far, you probably already know that BigPanda intelligently clusters your noisy alerts into high-level incidents. What you may not know, however, is that we have a Sharing feature, which makes it easy to notify and collaborate with anyone on your team about critical incidents.
So, if one person on your team is investigating an incident with an AWS-deployed app – an app that you’ve also ported to some other public cloud or to your own on-premises infrastructure – you can easily communicate via Text, Email, JIRA, and ServiceNow with your counterpart who is responsible for the other deployments.
Sharing an incident is easy and intuitive. When hovering over an incident, simply click on the Sharing button, and choose whomever you want to communicate with. That’s it.
You can share an incident multiple times and with multiple people, and the sharing history is logged, as well. Each outgoing Share also includes a link to Live Incident Information, which has up-to-date information on incident status. That way, anyone on your team can simply click on the link to get a real-time update on that incident.
As apps continue to sprawl anywhere and everywhere, DevOps teams need the same sort of agility that app developers have. With features like BigPanda Share, we’re helping DevOps teams catch problems in real time and remediate them across infrastructures. This is what’s needed to keep up with today’s rapidly changing app infrastructures.