To Build or Buy? Five Factors in the Cost of Log Management
Speed Drives Success for Cloud-Centric Businesses
Loggly customers are a smart bunch of developers, DevOps gurus, and CTOs who are taking full advantage of the speed and economies of scale that cloud technologies bring to their businesses. They know that the key to their success is moving fast and generating scale before the competition does. This means it’s critical to keep their engineers laser focused on innovation that matters. In spite of their need to focus on core competencies, our survey data tells us that one-third of our customers started trying to build their own log management solutions before deploying Loggly.
In this post, I’ll talk about five factors that you need to consider when trying to decide whether building and maintaining a homegrown log management solution is for you, or whether you should leverage a log management service like Loggly.
Factor #1: Log Data Is Unpredictable
Your applications log at a pretty steady state most of the time, until you experience a spike in your normal business or there’s some kind of fire. If errors and exceptions are spiking—third-party services are timing out, database queries are lagging, or end users are seeing a barrage of 503’s—your applications are probably emitting logs like crazy. Of course, those are the times that you need your logs the most.
At Loggly, we see customers generating more than four times their normal log volume, with these event bursts sometimes lasting several hours. If you’re building your own log management service, you’ll need to build it with lots of excess capacity if you want to be fairly sure that you won’t lose critical logs at the most critical times.
Factor #2: Log Management Demands Redundancy
Your log management service is of no use if it’s not available to you when you need it. Unfortunately, most internal logging services have no redundancy whatsoever. The end result is missing log data and days spent troubleshooting problems that could have been solved much faster with the right insights.
Factor #3: Proactive Log Analysis Requires More Data Retention
Logs are valuable for more than in-the-moment troubleshooting. Typically, as organizations mature in their DevOps practices and begin to use log data strategically and proactively, they recognize that they must keep this data for longer periods of time—increasing the hardware that you must deploy and manage. But more significantly, systems that work well with small amounts of data must often be designed entirely differently when the data volumes grow by a factor of 10 times or more.
Factor #4: There Is a Cost to Keeping Log Intelligence in the Hands of the Few
Unless you have some amazing developers on staff with time to spare from your core product, ease of use will probably not be a strength of any homegrown log management system. As a result, any root cause isolation efforts tend to funnel through one or two people who have conquered the learning curve for querying data and manipulating results. A glamorous job? Not exactly…especially at 3 a.m.!
Having a single source of truth that’s accessible to everyone on your development and ops teams makes everyone happier and more productive. Some Loggly customers, like Segment, have even been able to open up log data to their technical support teams. For Segment, extending log access means that developers need to get involved in only the most serious troubleshooting problems and can focus more of their time on innovation.
Our customer xsolla saves three person-hours per day because of Loggly, and other customers have reported similar results. Most importantly, accessible log data facilitates the kind of transparency and communication that are hallmarks of a DevOps culture.
Factor #5: Log Management Represents a Huge Opportunity Cost for Your Developers
Log management is hard to do well. We know because we’ve been in the business for more than five years. And unless you’re in the business of log management, chances are that the skills you’re building in log management don’t translate all that well to the software product that keeps the lights on at your company.
Every person-hour of development that you invest in log management is a person-hour that you aren’t investing in features that make your own solution special. If you’re not adequately budgeting for the costs and time of homegrown log management, the ripple effects can be huge in terms of lost productivity and time-to-market delays.
Keep in mind that developing your log management solution is only one piece of the puzzle. Operating a homegrown log management service is no different than any other service in your application; you must have someone whose job is to keep the system up. The tasks are fairly simple at low log volumes, but at scale you will be managing multiple servers and performing index/node/cluster management. These are not simple Linux sysadmin jobs. Even if you have someone with these skills, that person will be spending 10% of his or her time doing something that’s not part of your core business.
You also need to consider training, ongoing code maintenance, and support. Even if you’re relying on open source software for the foundation of your solution, all of these other costs will apply.
A Log Management Service Delivers Cost and Agility Advantages
Rumble Entertainment, developer of free-to-play online games like KingsRoad, used to run a homegrown log management solution. According to Albert Ho, executive producer and product manager for platform, “Getting good at log management wasn’t going to make us better game developers, so we were better off finding a service to take it off our developers’ hands and focusing on our core competency.” Drilling into the economics, Rumble told us that replacing its homegrown log management solution with Loggly doubled its log management capacity, tripled data retention time, and still reduced its monthly costs by more than 75%!
Jonathan Keith is the development manager for Monex Insight, a financial web application serving about 1.5 million consumer investors in Japan. His team adopted Loggly out of concerns about the resources that went into maintaining its homegrown log management at the expense of other deliverables. “It was a constant thing that we had to think about,” Keith explains. “The time updating servers and worrying about disk space was time we thought would be better spent focusing on our core competency.” He estimates that his team gained more than five person-hours per week by moving to Loggly.
Conclusion: A Log Management Service Focuses You on Your Core Differentiation
If you want to keep your application healthy and your company’s revenue flowing, you need an effective way to use log insights to solve operational problems, monitor performance, and inform development decisions. It’s tempting to try to own every aspect of your service, but the biggest lesson that the cloud revolution has taught us is that agility and elasticity count for a lot. So stay focused on your core business and leave log management to us, your resident experts.
Want to see how these five factors play out for your organization? Sign up for a free trial of Loggly!