Two months ago, I wrote about making CICS and Db2 data via OMEGAMON data provider
here.It took a bunch of effort, but sample dashboards for each have been posted
here.
What’s New with the Dashboards?
When the z/OS dashboards were originally developed, our team didn’t think as much about the scale and usability as much as they did to deliver visually attractive dashboards. Development of the CICS and Db2 samples identified some issues that had to be rectified. The results, something better than originally anticipated.
ODP has adopted a style guide for sample dashboards. The look, feel and navigation for each dashboard is similar across all the subsystems.
The Data Inventory dashboard turned out to be very expensive to operate if all subsystems were included, but it was also useful to see what attribute tables were being transmitted to ODP. As a result, the data inventory is done separately for each subsystem now and is accessible as a link on each dashboard.
New Documentation
The Installation and User’s Guide does an excellent job of describing the z/OS and Java server parts of the offering. Initially, there was very little on the Analytics and set up for those environments. That’s all changed. A new website,
https://z-open-data.github.io/odp-elastic-samples/ has been created with detailed documentation on the customization and set up of the analytics environment.
Hints and Tips
Codepage normalization
Customer experience has us constantly learning as well. One thing is handling code page variations that deal with ASCII related content used by the ODP Data Connect server in JAVA. In particular, the square bracket characters “[“ and “]” are not common across EBCDIC code pages and translation issues created operational difficulties. We found this because the 3270 terminal emulator may be operating with a country specific code page, the ISPF editor in another and the data being represented in still another. This was particularly bothersome in a few European countries with different national code pages.
As a result, we’ve changed our examples, as much as possible, to adopt the list form of notion instead of the bracket form to avoid these incompatibilities.
- open- pds
Elasticsearch Index
Another change is around Elasticsearch indices. You should adopt an Elastic Index Lifecycle Management (ILM)
policy. The original recommendation was to include the timestamp in the index. Net result was far too many unique indices which created a scale issue.
New indices will leverage the attribute table name instead to reduce the volume and simplify processing. The ILM can reduce the number of indices over time.
Logstash pipeline definitions
ODP originally included an 18 line logstash file for processing ODP attribute tables in a generic fashion. This works well when OMEGAMON is the only source of information to the Elastic environment. However, if OMEGAMON is being added to an existing Elastic environment, with other data, then an
updated logstash file that uses variable substitution can be utilized to ease the merger of data sources.
Because of the above changes, it is recommended that existing customers take advantage of the new dashboards and index files by updating via the code supplied within GitHub here.
What’s the future hold?
Soon, IMS and JVM attribute tables will be available via ODP. In addition, some functional updates will be made to further enhance the usability. Stay tuned.
Other useful information.
Give it a try. Dave Willoughby at IBM has been having fun building mashups from Log and OMEGAMON data. You can read about his experiences
here.
I’ve done a nine-minute video as an overview of ODP available
here.
Don’t hesitate to reach out if you’d like to learn more about this environment and how your business can benefit.