By Eric Bruno
One of the unintended side-effects of the public cloud is the enablement of a platform-agnostic cloud architecture. This means that it’s now easier than ever to mix offerings from multiple vendors (cloud vendors in this case) to form a single application or service-based solution. There is one gotcha, however: log management.
What Is Hybrid Cloud management?
A hybrid cloud links on-premise infrastructure with either a private cloud (managed internally) or a public cloud (managed by a third party). Even though you could say that private clouds and public clouds fall under the same category, they are completely unique infrastructures. Hybrid cloud solutions allow for organizations to host their platform or framework on multiple deployment environments which can be beneficial in many ways.
Benefits: Why choose hybrid cloud management?
Data Continuity
Hybrid cloud solutions are regarded among organizations of all sizes as an important part of their business. Hybrid cloud solutions allow companies to avoid any downtime and costs that would cause. This provides “data insurance” in the event of unforeseen downtime. Since the solution is to link public cloud, private cloud or an on-premise solution, your data is available in several places, allowing you to troubleshoot much faster.
Risk Assessment
For those companies who are just now looking into the cloud, whether it’s managed by them or managed by a third party, Hybrid cloud solutions allow companies to test their transition to the cloud before actually doing it. These companies can move parts of their servers to the cloud and keep some in their servers on-premise, making the transition smooth over a longer period of time, versus having to make the transition to the cloud all at once.
Quicker Deployment
In today’s market, we all know that a company’s ability to deploy, test and launch a new product depends greatly on the environment that their infrastructure lives in. Hybrid Cloud allows companies to test different environments and improve the speed of deployment. With multiple deployment environments, companies can launch their products faster and gain a major share of the market.
Market Share: Hybrid Cloud Adoption
The growth of hybrid cloud began in the early days of public cloud adoption. This offshoot was often called ground-to-cloud deployment, or the mixing of on-premises software deployments within private data centers and single public cloud service usage. For example, the use of in-cloud development and deployment tools, or cloud-based storage and associated tools for logs management and analysis, were two common hybrid cloud use cases early on. In both cases, early cloud adopters left the key components of their application architecture in their data center.
As use cases have proven that security and performance risks can be mitigated, and that true cost savings are there, the use of cloud services has grown. At the same time, the use of cloud services as a form of shadow IT has also grown, where line-of-business owners circumvented their IT constraints to become more agile. Although it was also possible with on-premises installed software, public cloud service usage of standards and web protocols make it much more straightforward to integrate multi-vendor offerings.
As a result, the growth of heterogeneous cloud service usage and tools has led to a new category of hybrid cloud: multi-cloud vendor-based solutions. To many, the true benefits of the cloud (security, elasticity, endless scale, reduced cost, becoming vendor-agnostic) can only be achieved through hybrid cloud usage. As public cloud benefits are proven through this approach, managing disparate cloud components requires careful planning.
Let’s look deeper at the challenges and potential solutions to hybrid cloud monitoring and management.
Common Hybrid Cloud Management Challenges & Considerations
The first step in meeting the challenge of hybrid cloud management is to focus on the essentials: system security, software delivery, and overall reliability. Often these are related, but require individual steps and solutions to achieve.
Here are the top considerations for implementing & managing a hybrid cloud environment:
Hybrid Cloud Security: Concerns & Vulnerabilities
Security and governance are a primary concern for those already in the cloud. You need to actively manage security, as with any application, but you also need to leverage cloud-specific management tools. This includes identity and access management (IAM) based security across multiple vendors, data security for both cloud-based data storage solutions (data at rest), and communications (data in motion). Management here involves monitoring encryption as data moves within and across services, while ensuring performance service-level agreements (SLAs) are met, and ensuring compliance when data residency and storage requirements are mandated, for example.
Managing Hybrid Cloud Infrastructure Deployments
Delivering software efficiently is an important hybrid cloud activity often overlooked in the application architecture phase for cloud-based solutions. This is equally true for enterprises migrating applications from on-premises to the cloud (where there’s a mix of both) as it is for applications that span multi-vendor cloud services. In both cases, the deployment of components to one area (data center or cloud) will vary compared with the other and even varies across vendors and tools.
One obvious solution is the use of management tools such as Docker and Kubernetes for container-based deployments. There are also many tools available to integrate on-premises deployments with cloud-based resources such as those for configuration management, automation in deployment and management, and deployment policy management to avoid risk.
Ensuring Service Availability
Hybrid cloud availability monitoring and management can be complex as it involves on-premises servers and cloud-based resources and data. Single view tools (often referred to as single pane of glass) are essential to reduce complexity and ensure you don’t miss something. You need to manage your own servers (their CPU load, overall storage capacity, network bandwidth, and so on) along with containers regardless of where they’re deployed. Specifically, containers bring their own challenges, such as monitoring the load and availability of resources within the container, as well as the physical hosting of the container while making sure container monitoring agents don’t adversely affect performance.
Tools, Strategies & Potential Solutions for Better Hybrid Cloud Management
As introduced earlier, hybrid cloud management requires deeper thought and planning than with other architectures. Because of the componentized nature of hybrid cloud design and the disparate services involved, it’s important to take a fresh look at how you monitor all the components that make up your application and manage them in case something goes wrong.
Traditional enterprise infrastructure management tends to focus on the components of software systems. It may include web and application servers, databases, and various network infrastructure components. Unfortunately, this siloed approach to measuring availability doesn’t give you a complete view of the user experience. It can indicate database server uptime or average transactional latency, but it doesn’t tell you about the customer experience.
Service Reliability
Service reliability monitoring builds on key application parameters of availability, but also measures the holistic user experience, ensuring that every user interaction with your software is successful. For instance, in a hybrid architecture, a user request may be processed by a cluster of on-premises servers while key portions of the request are also fulfilled by cloud-based services. Let’s assume that either a single on-premises server, network link, or cloud provider has just a slight issue. While traditional monitoring tools may show 100% reliability for both sets of resources, a deeper analysis is required to get a true picture of what the user is experiencing: a non-optimal experience. Learn how to measure your users’ satisfaction with Apdex Scoring.
The risk to your business is real: with 1 million requests per day, a reliability issue that affects as little as 1 percent of your users translates to 10,000 human beings who have had a negative experience with your application or service. To get to the true picture of what your users are dealing with every day, you need to monitor at a deeper, end-to-end level.
Make sure you watch and manage everything, including your local servers, integrated cloud services, container deployments, databases, user request-response transactions, network infrastructure, virtual machine, and language platform health, and software asset inventory for mismatched deployments.
User Monitoring
To further ensure an optimal hybrid application experience, you need to go beyond infrastructure management, where servers and IT resources are monitored and include monitoring from the end-user point of view. Look for management tools that detect anomalies and immediately alert you when the end-user experience is affected by application performance.
You need to inspect the details of each and every individual user session, tracking requests, response times for database queries, Apdex score, web page data aggregation, JavaScript user-interface reaction times, and cloud resource latency. Tools should proactively alert you to potential issues before your users notice, supporting dynamic threshold management for a variety of performance metrics. The key difference from a typical measurement strategy is to look at how a human being experiences a hybrid application.
Tracing the Transaction Envelope
True application availability is measured by analyzing each user transaction, from initial request to final response as viewed by the user, for each user transaction. For instance, if a specific transaction typically takes less than a second to complete, but with an occasional outlier, the entire transaction should be marked as suspect.
Get full application health visibility by end-to-end tracing of request execution across multiple applications, tiers, servers, microservers, and processes. If there are 10 components to every transaction that need to be monitored, but one of them is missing or out of range for a single transaction, the workflow experience should be marked as suspect.
Your management tool needs to analyze deep into application layers, from database operations including server performance and individual SQL query times and HTTP page construction from the application server through dependant services to language platforms such as Java virtual machine (JVM) bytecode instrumentation.
Careful analysis of performance metrics include:
- Latency: response time tracking for a single request or transaction
- Throughput: the measurement and tracking of average response time, i.e. requests (or transactions) per second.
The difference is key, as just monitoring average response time covers up the outliers. For instance, even when average response times can meet your SLAs, you can miss the few outliers that occur that adversely affect real users. Seeing both visually and in real time shows where in your hybrid architecture response time is being impacted.
Hybrid Cloud Log Management Tools & Solutions
With hybrid cloud, another challenge is log management, as components spread across vendors and locations may maintain their own logs. Log data is critical both to diagnosing problems post-mortem and to real-time user and application analysis. Capable log management tools allow tracking and alerting across multiple users, and hybrid cloud infrastructure metrics.
Tools such as Sematext Logs allow you to centralize and analyze your logs using your existing tools, helping to correlate across time and user requests. Then you can perform key log analysis, extracting key performance indicators and server metrics. You can also perform queries, set thresholds, and visualize the results with dashboards and reports. Further, log analysis is integrated across a wide range of tools, including email, Slack, WebHooks for cross-cloud integration, and more.
Use Sematext to gather all of your log data in one place securely and accurately, integrated with other tools through standards such as the RFC-3164 message format or JSON, and filterable and searchable using boolean operators and other search tools.
Trends
In the past few years, many organizations have adopted hybrid cloud environments, mainly due to the many benefits it can offer. Gartner recently made a projection estimated that the public cloud services market will reach $206 billion in 2019, which is a $30 Billion jump since 2018.
“The worldwide public cloud services market is projected to grow 17.3 percent in 2019 to total $206.2 billion, up from $175.8 billion in 2018.”
– Gartner, Inc.
Gartner has commented on the massive adaptation to hybrid cloud environments by many companies in the past couple of years. Hybrid cloud environments allow companies with on-premise solutions to take advantage of cloud-based offering, without having to throw away their investments in their on-premise solutions.
IBM has made some strides in the hybrid cloud industry with their acquisition of Red Hat in 2018, however, it doesn’t change that Amazon AWS and Microsoft Azure will continue to dominate the market. Leading with a 41.5% share of the market was Amazon AWS, Azure following with 29.4% of the market share. (Taken from Cloud Security Alliance (CSA) report)
Hybrid cloud environments offer the huge advantage of Disaster Recovery (DR). Using the cloud as a backup for data allows businesses to leverage the advantages of low-cost cloud storage as well as providing an offsite backup copy.
Wrapping up
Together, infrastructure management (both on-premises and in the cloud), real user experience monitoring, transaction tracing, and log management ensure a positive hybrid cloud application experience. This approach will help your organization realize the full benefits of the cloud through hybrid data and application architecture.
Bio
Eric Bruno is a writer and editor for multiple online publications, with more than 25 years of experience in the information technology community. He is a highly requested moderator and speaker for a variety of conferences and other events on topics spanning the technology spectrum, from the desktop to the data center. He has written articles, blogs, white papers and books on software architecture and development for more than a decade. He is also an enterprise architect, developer, and industry analyst with expertise in full lifecycle, large-scale software architecture, design, and development for companies all over the globe. His accomplishments span highly distributed system development, multi-tiered web development, real-time development, and transactional software development. See his editorial work online at www.ericbruno.com.