TWSGraph

1
TWS/Graph
Status Monitor
How to use the monitor
HORIZONT Software für Rechenzentren
Garmischer Str. 8 D- 80339 München Tel 49(0)89
/ 540 162 0www.horizont-it.com
2
TWS/Graph Status Monitor
This presentation shows you four examples how to
use the monitor

• Create job lists
• Monitor milestones
• Generate tickets (BMW)
• Monitor if production is late (Caja Madrid)

3
Example 1 Create Lists

• The first example shows how create easily job
  lists with the monitor

4
Create HTML Lists
This (simple) example shows you how to create a
homepage that meets following requirements

• List all jobs that are waiting
• List all jobs that are active
• List all jobs that are complete
• List all jobs that are in error

5
Create new file

• Create a new monitor definition

6
Insert top level

• Insert top level (any text is ok)

7
Insert ERROR object

• Insert Operation object to create ERROR list

8
Create other objects

• Create Operation objects for WAITING, ACTIVE, and
  COMPLETE

9
Arrange objects
New with V5R2 Order can be changed
10
Save definitions

• Save definition, then start

11
Start monitor

• Specify all needed HTML options

12
Review status
Monitor is now up and running. No errors!
13
Resulting HTML
Top level entry point...
14
Resulting HTML
Access to lists
15
Resulting HTML
PS You can use the link to access e.g. ERROR
lists directly. As long as you use this
definition, the file name doesnt change!
16
Resulting HTML
Detailed job information for all jobs that are
complete. Automatically updated every five
minutes!
17
Modify list layout in TWSGRAPH.CFG
Modify TWSGRAPH.CFGLIST_POSCPO_OP_contains
the names of columns
Remember All variable names are listed in
TWSGRAPH.LBL
18
Modify Sort order in TWSGRAPH.CFG
Change HTML Sort Order
19
Example 2 Milestones

• The second example shows how to use milestones to
  monitor production

20
Milestones
Probably the best way to monitor the Critical
Path is to insert milestones
21
Create Dummy Workstation

• Create a dedicated dummy workstation to be used
  only for monitoring

22
Create Dummy Workstation

• Create milestones as successors to important
  jobs

23
Create Dummy Workstation

• Specify deadline

24
Production with Milestones

• Milestones are now inserted into production
  (without any functional effect to TWS)

25
Create Monitor Definition

• The status monitor of TWS/graph needs only one
  object

26
Start Monitor

• Start the monitor

27
HTML with Milestones

• The resulting list shows if you are in time

or not Status L(ate)
28
Pros and Cons
We think Milestones is the best method to monitor
thousands of jobs

• Simply to use
• Excellent performance (TWS/Graph must monitor
  only the milestones, and not the whole stuff)
• If there is a known delay to the start of batch
  by, say 2 hours, you only need to change the
  'late time' field to '0200' in the status
  monitor definition
• Milestones must be maintained

29
Example 3 BMW

• The third example shows how BMW uses the monitor.
  

BMW has its headquarter in Munich. BMW is using
TWS/Graph, TWS/Audit, TWSBatchAD, TWSBatchCP,
SmartJCL and XINFO.
30
Initial Situation

• Some critical jobs where checked manually
• No checks between 0000 and 0500
• Tickets where created manually

31
Target

• Generate Alerts and Tickets automatically
• For Error Jobs
• For Late Jobs
• Generate Status Information
• For the programmers who are responsible for the
  tickets
• For non IT-departments

32
Status Monitor Web Designer
BMW is major test site for the new web
application to maintain the monitor definition.
33
Status Monitor Web Designer
TEC and Remedy parameters can be maintained
together with TWS/Graph definitions
34
File Export to TWS/Graph
In the actual test installation TWS/Graph Web
exports monitor definitions to TWS/Graph Client
35
Refresh Rate
The refresh rate is 5 minutes. One refresh
takes 3-4 seconds and requires 0,1 CPU seconds.
36
HTML Status Information via Intranet
The HTML directory is shared, anybody (who is
authorised) can view the files
Department names are used as entry points
37
Alerts and Tickets
In case of Error- or Late jobs a Perl script is
started
38
Perl Script

• TWS/Graph passes TWS, TEC, and remedy Parameters
  to the script

• The Script generates and sends Command postemsg
  to the TEC

39
TEC Events
TEC-Events generated by TWS/Graph
40
Remedy Tickets
Remedy Tickets generated by TEC
41
Example 4 Caja Madrid

• The fourth example shows how Caja Madrid uses
  TWS/Graph.

Caja Madrid is a Spanish bank, the headquater is
in Madrid. Caja Madrid is using TWS/Graph,
SmartJCL and XINFO.
42
Initial Situation

• Tickets are generated automatically (EQQUX007)
• They have few but important key applications
• The Batch Window is too small
• Operating is not able to see possible delays in
  time

43
Target

• Operating needs early warnings when a critical
  job will probably not end in time

0600
0500
44
Technical Overview
Based on Caja Madrids requirements, HORIZONT
developed a monitor concept

• Only critical end jobs must be defined to
  TWS/Graph
• TWS calculates Planned Duration for every job
  while extending the Current Plan
• TWS/Graph searches all predecessors of the
  critical jobs
• TWS/Graph uses the Planned Duration to find the
  critical path for every critical job
• TWS/Graph calculates the estimated execution time
  of every critical path and shows the remaining
  buffer time

45
Jobs on the critical path(s)
TWS/Graph can monitor as many critical end jobs
as needed. Here we have two critical end jobs.
Das kann jetzt (5.1) auch der Status Monitor for
einen oder mehrere Ende-Jobs
46
Late Options
Use the TWS/Graph dialog to define an object for
every critical end job
The critical job (Application Id and Operation
Number
Aaa
Late time as an offset to IAT
Predecessor levels that should be taken into
account
47
Calculated remaining duration Path
The end job and all predecessors that are not
complete
After having started the monitor, TWS/Graph
automatically builds a flowchart with all jobs
that are on the critical path
48
Calculated remaining duration
Information window
Number of jobs on the critical path
Calculated Late-Time 1700
Actual buffer1 hour
Status
Remaining runtime 3 hours
49
Calculated remaining duration
The actual buffer is calculated frequently
1330
50
Calculated remaining duration
If it looks like the critical end time will be
missed, the buffer becomes negative
1401
51
Calculated remaining duration
The Monitor signalises Late
52
Calculated remaining duration
Jobs are running now, some jobs are complete
Now the buffer is sufficient again
1407
53
Calculated remaining duration
Path becomes shorter
1430
54
Calculated remaining duration
Summary TWS/Graph can show you where you are and
the best estimate of when the Batch will end
1500
55
Some figures
Even we where concerned about performance, but
the result was better than expected

• 6 critical jobs are monitored
• Refresh-Rate is 1 hour
• Runtime (one Monitor for 6 Jobs)
• 18 Minutes right after Begin of production
• 10 Minutes just before End of production
• Runtime (6 Monitors parallel)
• Each monitor requires 5 down to 2.5 minutes

56
Benefits of TWS/Graph Status Monitor
Let us summarize the benefits of TWS/Graphs
integrated Status Monitor

• Easy to install (no installation if you have
  TWS/Graph)
• Different targets can be addressed (Lists,
  Tickets, Late)
• No system modifications needed (Exits etc.)
• All jobs controlled by TWS can be monitored, z/OS
  as well as End2End jobs

57
Do you have any questions?
Support_at_Horizont-it.com

• Create job lists
• Monitor milestones
• Generate tickets (BMW)
• Monitor if production is late (Caja Madrid)

58
Many thanks for your attention!
TWS/Graph
Status Monitor
HORIZONT Software für Rechenzentren
Garmischer Str. 8 D- 80339 München Tel 49(0)89
/ 540 162 0www.horizont-it.com