A key tenet in the DevOps movement is to improve the communication flow between Development and Operations. This is achieved by embedding Dev and Ops into each other’s processes at every step in the application life cycle. The desired result is a shift left of problems—finding issues much earlier in the life cycle when they are cheaper to correct.
For decades, Abend-AID, Compuware’s mainframe application failure resolution and application fault management solution, has lived at the intersection of Development and Operations; traditionally on the far right of the application lifecycle, used to troubleshoot problems in production.
Abend-AID has always earned its living by reducing time to resolution at the most expensive point in the cycle—when problems are affecting production systems and customers. In many ways Abend-AID is the lingua franca between the two cultures: showing issues from an Operations perspective with background detail needed from the Development perspective.
But Abend-AID provides so much more than that. Unique in the industry, Abend-AID captures metadata about problems occurring in your systems (development, test, production) to allow sites to begin a true process of mainframe fault management and to shift problem detection to much earlier in the application lifecycle. This capability provides significant benefits:
- Detects Issues Before Production
When used in development, analysis of this metadata allows you to detect worrisome trends before they affect production systems and your end customers. Consider it Newton’s fourth law of Motion: “Applications abending tend to keep abending.”
- Acts as a DevOps Juncture
It gives Development and Operations a common meeting point: the ability to determine if mainframe IT quality is improving or regressing; the ability to categorize problems as environmental or programmatic; the ability to deal with problems as groups to get the most benefit from coding efforts. The flip side of Newton’s fourth law: “Applications that don’t abend tend to keep not abending.”
Let’s run through a quick example of this metadata. Abend-AID provides a component, Fault Analytics, to process this data, but you can also feed it into other analytic engines if you desire. In this case let’s look at that data through Splunk.
This Splunk app breaks down your fault information (production or development or both) into six areas of information. The left side shows recent history; the right side shows today. The top shows what is happening; the middle shows where errors are occurring; the bottom helps you determine why those errors are occurring. Every pie slice/bar is clickable to see the supporting data underneath it, which can then be summarized to drive deeper analysis. This information can be summarized as:
This provides some at-a-glance analysis:
- What is the sheer volume of abends occurring on this system? Is there any trend? Is today noticeably better or worse than the most recent past?
- What is the personality of these abends? On what machines and which subsystems are these abends occurring? Is today consistent with the patterns shown in the most recent past?
- What is the key underlying information on these abends? Is it the same abend occurring over and over? Are the abends occurring on recently changed code? And again, is today consistent with the patterns shown in the recent past?
This capability to take a step back and analyze the forest for the trees provides fertile common ground for your organization. Development and Operations working together, using a common language, to begin to manage problems as early in the application life cycle as possible; to improve the quality of production systems; and, ultimately, to better service your end customers.
Long before the term DevOps was coined, Abend-AID facilitated the interaction between Development and Operations. And, using Abend-AID’s industry unique fault metadata analytics, you can leverage Abend-AID to shift mainframe problem resolution left in the DevOps lifecycle.