Welcome Back
Welcome back to the blog series on the OMEGAMON Data Provider (ODP). If you've missed the excitement about ODP, you'll want to read the prior two blogs, and definitely sign up for the upcoming webinar on 2/22
http://ibm.biz/AIOps-022222 !
Endless possibilities with great multi-source mixes!!
Everyone loves a good music mix! While individual artists have awesome music, I've found that the relatively new music mix phenomena is so enjoyable -- its great how some of the techno artists can draw multiple sources into a new creation all its own!!
System Administrators, Site Reliability Engineers are stretched. When issue occur, there's a lot of disparate logs, reports to peruse, to get a sense of what's going on. Even more so, veterans have a knack to know what's best to monitor before issues occur, be that metric values or specific log messages or both. I've had customers tell me they religiously monitor the full round trip, from front end to back end, and back to the front end, and just monitor the request rate and average response time. When these end-to-end SLAs change somewhat, then they look deeper. When they look deeper, i'm betting they'd love some great "mash ups" of various logs and reports, to quickly look deeper.
Laying down "First Tracks"!!
I recently started looking at easy log and metric mashups, that require no code, are simply visualization engineering!!
I'm sure i'm just scratching the surface. My first mashup used SYSLOG and ODP Address Space CPU metrics, where I put both on the same Kibana ELK dashboard, after streaming both into the same Elasticsearch via Logstash. The SYSLOG part of the dashboard does a keyword search of SYSLOG, searching for JOBNAME.keyword, where JOBNAME.keyword is the top CPU address space JOBNAME! This was so easy and straight forward to create. SYSLOG was streamed using IBM Z Log and Data Analytics (IZLDA), and the Address Space CPU metrics were streamed via ODP.
Here's the result, a simple log and metric dashboard, so cool!
As I said, I think this is just scratching the surface, there are limitless mashups that are possible. I think one of my next mashups will use application "trace" data, which gives a "vertical" view of a single end to end application thread, while SYSLOG and ODP give a "horizontal" view of various layers of the underlying OS and Middleware.
What mashups are you thinking of? Comment below!!!