SmartTrace

Smartest way of TIBCO Monitoring

SmartTrace

We are proud to announce SmartTrace - a light-weight but powerful monitoring product that monitors TIBCO Businessworks process durations and counts with their backend urls.
(HTTP, JDBC, JMS, FTP backend response times)
SmartTrace also monitors EMS traffic and pending messages and the undelivered queue ($sys.undelivered)


Questions

Questions

Common tough questions on an ESB middleware layer:

  • What are the backend urls being consumed by the ESB layer and their response times and transaction counts for every minute period? How to visualize them?
  • What are the entry points (start urls) of the ESB layer and the response times of each entry point?
  • How to make trend analysis on the outgoing and incoming traffic of ESB layer?
  • What was the trend (response times and counts) of a specific operation on ESB for the entire week/month (eg. Mobile Banking Login trend for the last week)?
  • Why did the response times increase at a specic date time on the ESB? What was the problematic backend that had poor response time?

With SmartTrace, it is very easy to answer all of these questions on a TIBCO ESB layer. A few clicks are enough!


Monitoring requests passing through businessworks engines

SmartTrace records every running process durations and counts inside a BW engine for every 15 secs periodically and discovers the backend urls being called. It learns the Average Reponse Time of a BW process and marks a higher(poor) response time at a specific time range in “Response Hit” column. For example if Mobile Banking Login operation normally takes 1 sec in average but it takes 3 secs at a specific dateTime then SmartTrace marks it as “3x” for that time period.

Date Time Filter

User can select a specific date time range to examine the requests’ response times and counts at that time interval.

User can filter and sort the requests by:

  • any date time frame
  • process/operation name
  • endpoint url
  • process type (Starter or Backend etc)
  • BW app/instance name
  • hostname
  • response times
  • counts
Figure 3: All Requests Screen
All Requests

Analysing a specific BW process response time and its backends

User can click on a specific process and do the following analysis:

  • choose a date time range and observe its response time graphically at that time frame
  • change Y axis of the graph to ‘Counts’ or ‘Errors’ or ‘Thread Consumption (ms)’ of that process
  • discover the activities and related backend urls of the process
  • Jump to another BW instance executing the same process and observe that BW instance’s response time
  • can see lots of information about the process like: the global variables, timeout values, SOAP WS operation name, SQL statements, etc.
Figure-4: Process Type Selection
Process Type Selection

SmartTrace discovers the processes inside the running BW engines and analyses them in a smart way. SmartTrace groups the running processes into 4 parts:

  • Starter process
  • Sub process
  • Backend process
  • Simple process

User can filter by these process types and observe their response times

Figure 5: Process Analysis
Process Analysis

Create your custom charts

User can define his own bookmarks and can bookmark any BW process, so can monitor all the bookmarked processes in one single chart.

For example user can bookmark processes from different BW engines making JDBC calls and compare them all in one chart

Create Custom Charts
Figure 6: JDBC response durations from different BW engines (Bookmarked Processes Screen)
JDBC Calls
Figure 7: JDBC request counts from different BW engines (Bookmarked Processes Screen)
JDBC Calls
Figure 8: Critical WebService calls invoked by different BW engines (Bookmarked Processes Screen)
Critical WS Calls

Creating and Monitoring Dashboards

User can create Dashboards and put N charts, so can have N different charts in the same Dashboard for monitoring.

Dashboards have auto-refresh feature and user can define the Auto-Refresh interval

Creating Pages
Figure 9: A Page of selected 4 bookmark charts
Monitoring Pages

Monitoring Endpoints: outgoing request counts and response times

All HTTP, JDBC, JMS and FTP backend endpoints are listed uniquely in Endpoints screen. User can easily address which BW process is invoking which endpoint url and what are the response times and request counts from each BW process to that endpoint

Endpoints
Figure 10: All Endpoints listed uniquely - user can expand an endpoint and examine the BW process durations invoking that endpoint
Monitoring Pages

Monitoring Start (Entry) points of the ESB: the incoming request counts and ESB’s response times

Every ESB layer have HTTP (SOAP) and JMS entry points. SmartTrace detects all the HTTP and JMS entry points (urls) that clients are invoking to reach TIBCO ESB layer and lists these entry points uniquely on the “Start (Entry) Points” screen. (See figure in the next page)

Start Points
Figure 11: Start (Entry) points listed uniquely - User can expand a start point and examine Starter processes for that entrance
Monitor Start (Entry) Points
Monitoring BW Heap

Monitoring BW heap usage and running job counts

SmartTrace monitors BW heap usage and running job counts of a BW engine.

User can browse to any time frame and analyse the heap usage and job counts graphically

Figure 12: BW engine monitoring (Heap Usage and Job Counts)
Monitoring BW engine Heap and Job Counts
Zero Affect

ZERO affect on the running BW app’s performance and %99 percent correct measures on response times.

SmartTrace is monitoring production environments that have 5000+ requests per second in various big TIBCO customers and many experiments proved that SmartTrace has ZERO affect on the running BW engine’s performance.

It does not affect the bw engine’s performance while monitoring. It is very light-weighted but very powerful with its smart algorithm.

Many performance comparisons made between BW engines connected and not connected to SmartTrace proved that SmartTrace does not put ANY extra load on a BW engine and does have ZERO affect.

Many production problem analysis made with SmartTrace in customer side proved that SmartTrace discovers backend urls and process informations %100 percent correctly and measures the response times and counts %99 percent correctly.

Outline of Major Features

Realtime Instant Monitoring

See last minute's process response times and request counts and browse any time frame and observe the response times at that specific time range!

15 seconds advantage

Get alerted in max 15 seconds after the incident! User can define email alerts based on conditions (increase in heap usage, number of hang or aborted processes, etc). Users can see all the alerts on Alerts window.

Intelligent Discovery

SmartTace discovers SOAP, JMS, JDBC, FTP backend urls inside the running processes at runtime and shows their durations for the last minute or a given specific time range. User can easily recognize which backend url is having long response time.

Artificial Intelligence

SmartTrace analyzes the processes and groups them according to their complexity and helps users to have better insight on the running processes. SmartTrace analyzes the problematic bwengine and suggests tuning recommendations (eg. analysis for root cause of hang processes).

BW Heap Monitoring

SmartTrace collects bwengine's heap usage statistics and number of running jobs in the memory and shows them to the users. Users can easily define alerts based on used heap percentage or number of running jobs.

Historical Analysis

Users can browse backward to do historical response time and bwengine analysis. SmartTrace helps users to recognize unused processes inside the deployed Ears (eg. never called processes since 1 year)

Easy to Use

Instantly start monitoring your BW engines in no time! It is very easy to add a new BW engine to start monitoring. Just a few clicks enough! SmartTrace has very user friendly dashboards and charts.

Extended Administration

BW developers can enable trace of a BW process and can download BW engine’s log file with just a few clicks at runtime. This helps developers to solve runtime (eg. UAT env) problems in a short span of time. BW developers will no more need BW admin guys frequently to solve problems. Note that only authorized users can enable trace. Admin guys can give 'Trace Authorization' to specific users.