Tracing Transactions from Mobile to Mainframe (Part 2): What you can see on the Dark Side of the Moon
In Part 1, Ian Hartley revealed how to illuminate transactions that have fallen off the radar onto the dark side of the moon. In this post, he looks at what you can see when you shed that light with Syncsort’s new product.
Seeing the Dark Side via HTTP(S)
With Ironstream® Transaction Tracing (ITT) you can:
- Leverage a low-impact, simple solution to track individual transactions across your infrastructure components and get information about their response times and health
- Monitor a specific system and/or action to see if that multi-tiered environment, including mainframe, is delivering good service to your customers
- See when a critical service starts to underperform and could take corrective action before your user community becomes aware
In the initial release of ITT, we focus on a popular option for mobile to mainframe transactions:
Clients using HTTP or HTTPS to communicate with CICS® via:
- Web Services (XML/SOAP or JSON)
- RESTful API
In upcoming releases, we will cater for other methods such as CICS Transaction Gateway®, calling DB2® directly (via DB2 Connect®, for example), WebSphere®, etc.
Below is a schematic of a typical http(s) transaction flow that can be monitored by ITT:
Figure 1. Schematic overview of Ironstream Transaction Tracing
A user interacts with a mobile application (UI) which makes a http(s) call to an exposed CICS web service running on the mainframe. There may be infrastructure components between the UI and CICS, but ultimately the request enters a CICS system.
The UI can optionally add a unique token into the headers of the http(s) call. Alternatively, ITT can generate a unique transaction token to be returned to the client.
When the CICS web service processes the call, it may, for example, interact with a DB2 database – both components will generate a machine data log.
Syncsort’s Ironstream, the high-speed, low-impact mainframe data forwarder, will capture these log records in real-time and immediately send them to Splunk.
Finally, the request returns a response to the UI. At this point, the UI can optionally log a record in Splunk to close the user experience loop. Without an explicit UI log record, other sources such as web server logs may be used to capture UI interactions.
Joining the Dots
Now log data exists in Splunk, the output records can be correlated. For example:
It is possible to:
- Calculate overall end-to-end user experience time (includes network latency etc.).
- Calculate time spent within each mainframe component.
- See details around each mainframe component.
- See if the transaction worked, or not.
- See if the transaction met custom performance thresholds.
Plus, in Splunk, we have a sample Ironstream Transaction Tracing app to show the mainframe portion of the transaction information.
In the implementation below, the Overview dashboard monitors systems and corresponding actions with average response times over a given time period:
- Mobile Bank
- Car Insurance
Figure 2. Ironstream Transaction Tracing overview dashboard.
This dashboard refreshes automatically and uses customizable thresholds to change a tile’s colour as the average response times are monitored.
Clicking a tile figure will take the user to a Transaction List dashboard where, based upon the time range, selected System and Action, a list of matching transactions, with the slowest at the top, is shown together with:
- An overall average response time
- A horizontal bar chart with time spent within each mainframe component
Figure 3. Ironstream Transaction Tracing transaction list dashboard.
From the transaction list, an individual transaction can be clicked to navigate to the Detail dashboard. Here, more detail is presented for the selected transaction:
- Timings between mainframe components
- Information from each component
Figure 4. Ironstream Transaction Tracing detail dashboard.
Further information could be gathered from other log data.
We could even reach out, for example, to the mainframe to gather more detail… or take some corrective action. An authorized user could be given specific options for solving a situation and redressing an imbalance in a facility that has started to underperform.
Plus, ITT can be bound into solutions such as IT Service Intelligence from Splunk.
With Ironstream Transaction Tracing it is now possible to shine some light onto the dark spots of your transactions as they travel across the infrastructure. They no longer drop off the radar.
It’s lightweight, non-intrusive mechanisms allow organizations to now see the complete end-to-end journey, including mainframe.
To learn more about Ironstream Transaction Tracing, watch our recorded webcast: End-to-End Transaction Visibility from Mobile Devices to Your Mainframe