Version: 22.61.37+
Move objects (primarily reports) with the tree structure
Another action for CMS DBx (Admin Tools 2.0)
Problem:
Thousands of objects need to be moved from different folders in which they are mixed with other objects (long time not refreshed, created by the same person, with specific name pattern, with a particular data source,…)
The move could be preferred over delete, as it allows further manipulation with objects, or they could be quickly restored by folder.
Features and usage:
-
Load objects that you are interested in with CMS DBx
-
Apply custom action to selected objects or subset
-
Actions can be easily added on demand
 |
|---|
Version: 22.58.36+
XML – compare and explore
Problem:
You need to compare two or more XML files (configuration, input/output,…) or make sense of what is in this vast XML.
It is easy to do with beautiful free line-by-line comparison tools, but what if XML is not ordered or too big, dozens of megabytes, or there are too many differences?
In this case, the line-by-line comparison takes too much manual effort and time or simply doesn’t work.
Features and usage:
-
Compare XMLs (not necessarily one to one, it could be one to two or three to five…) by opening them in BiWhy
-
Hide common entries
-
Ignore duplicates (XML node with the same path/name/value) or not
-
Hide empty or zero valued nodes
-
Show short paths in table view when possible for readability
-
Delete XML nodes from analysis for readability and performance, if you are not interested in them
-
Expand/collapse branches wide or deep wise
-
Filter by XML path
-
Convenient backward navigation from table to tree location
Example:
Compare service.xml to a few other versions
 |
|---|
Version: 22.57.33+
Files (log, dumps,…) collector
Problem:
Log collection is a dull, time-critical, and time-consuming process
Frequently logs need to be collected in production as an issue is not reproducible in lower environments
Production has multiple hosts, hence logs need to be collected from multiple locations
Logs need to be collected quickly as production usually generates tons of noise (NB: BOBJ admins and developers need to address this)
Workaround – Creating a server group, is by itself time-consuming and not always possible
Workaround – Increasing logs retention count, is also time-consuming and leads to dozens of additional gigabytes to be transferred and analyzed (frequently such volumes are even hard to open)
Manual collection easily can take an hour
Collection might need to be repeated multiple times
Features and usage:
CMS and Processing nodes’ log locations are automatically generated
Tomcat template log locations (four different places). You only need to update hosts and user in it
Overall configuration takes only a couple of minutes and is only needed once
Filter by last modified and file extensions to concentrate from the beginning on relevant data and minimize collection and transfer time
Report source folder volumes (size and quantity) – could be circumstantial evidence for problems on a particular host or problems with load balancing
Collection is running in parallel threads, minimizing collection time
Report transfer progress and speed
Log transfer errors
Wild card support in source locations
Further development for wider usage: recursive flag, preserve relative path locations
Example:
A system with ~dozen nodes and few Tomcats resulting in ~30 log locations
Recent glf log collections take only a couple of minutes
Collecting all logs takes ~20 min. Of cause, time depends on the number of logs, network/disk performance, and number of CPUs, which is in times or even in order of magnitude (for recent logs) is quicker than manual collection.
Inconsistency in the system installation was found just based on Tomcat log locations (reported absence)
 |  |
|---|
Version: 21.53.32+
CMS DBx (AdminTools 2.0)
It is a long-awaited and very versatile tool.
Problems:
-
Query CMS DB:
-
CMS DB is encrypted and can't be queried by conventional SQL tools.
-
BOBJ SDK (AdminTools) doesn't support full ANSI SQL, partly compensating it with proprietary syntaxis. Still, many things can't be queried.
-
AdminTools output is a tree-like table HTML, which is unsuitable for efficient analysis, especially for bag stored properties.
-
No simple interface to perform custom actions (delete, enable, reset passwords…) on a custom set of objects from the above query.
-
No simple interface to run RESTful calls.
Features and usage:
-
Query CMS DB from BiWhy and get results in an automatically generated object tree or table.
-
Use one of more than 20 templates (growing list) for a quick start or learn about BOBJ-specific syntaxis. Where required, templates are provided with comments and links for further readings.
-
Filter in place and analyze results in tree/table/char views.
-
Convert to a flat table and export to excel (in the process).
-
Interface allows quickly adding custom actions for one or mass object processing. We will be adding them on demand.
-
RESTful actions.
-
Derivative properties, calculated on the fly. Example: Duration for WEBI reports.
Example:
A customer wanted to fetch a list of disabled users. They were doing it manually, as they haven’t found a way to filter it in AdminTools.
The process took hours (small system and there are systems with hundreds of thousands of users).
With BiWhy it took only a few seconds, without any development.
Version: 20.46.30+
CMS DB, file statistics switch
It is a slight improvement, but File Repository performance is often neglected.
With the File Statistics flag, you can compare the performance of CMS DB metadata fetch by a few indexed fields and the same query plus file information (FRS calls and unindexed fields).
In a few extreme cases, fetching file statistics took hours or even timed out, though overwise CMS DB, based on Average Query time, looked utterly healthy.
In contrast, it takes only a few minutes on quick systems with comparable CMS DB size.
Don’t forget run statistics collection on CMS DB periodically. Surprisingly, it helps.
Version: 20.41.29+
GLF Logs Analyzer
There weren’t updates recently in “what is new,” but not because there is nothing new; on the contrary, we were swamped developing long-anticipated features.
Here is the first, but not even the best new one.
Several improvements in the E2E (end to end) GLF (BOBJ's log formant) analyzer:
Short Critical Path Tree – workflow to key time allocators with only essential forks in the path.
Noise filters – allow hiding verbose logs to grasp main workflow branches and improve chart view performance.
Find – search in visible fields, and filtered results view.
Flat stack view.
There are several minor improvements to see time allocation in tables or charts distributed by Host/Serve Type/Server/Thread/Action/Caller Function/Callee Function/User.
Version: 17.40.25+
System Health Checks
BIWhy analyzes your system metrics and triggers Red/Yellow/Green alerts when applicable, based on SAP documentation and experience accumulated by BIWhy
Metric are accompanied by extracts from SAP documentation or/and by BIWhy's comments.
Checks are triggered by expected thresholds, when known, or arbitrary thresholds to attract attention to magnitude when there is no sharp edge between healthy and problematic areas.
Other checks are triggered by expected values or by values comparison between servers (processes in BOBJ terminology) and hosts.
Though in some cases Red/Yellow/Green means problem or its absence, in most cases they are rather used to attract attention to potentially problematic areas.
System Health Checks Overview
System Health Checks Overview
Version: 16.40.25+
Metrics Grouping to quickly find one you need, see what is available in this area and what are it's values for different Servers.
WebiServerAdmin Metrics
5min CPU Metric from JAVA family
WebiServerAdmin Metrics
Version: 16.37.24+
Java Thread Dumps Analysis
Analyze dump sets by CPU, Input/Output Disk/Net IO, opened files or sockets.
Drill to problematic dump set, thread groups, thread.
Find what thread is doing by browsing tread stacks, metrics correlation, state or statuses over the time change.
Version: 15.36.20+
Build Landscape from collected Metrics.
Hosts with their configurations (CPU, RAM, OS and BOBJ Versions, Disk).
Immediately see how summary of Memory or CPU consumption, from reporting Severs, relate to available resources.
Configure what you want to see.
See if configuration does not exceed hardware limits.
Check if parameters are set to the same values across all Servers.
Version: 14.34.19+
Added Publication Charts and Filters
