This document provides a high-level summary of the UI and performance improvements implemented in 4.2.1 Release along with some other features. We assume that below enhancements will improve the overall performance of Cavisson products.

NetStorm

TestSuite / Jenkins Enhancements

‘Overall Fail Criteria’ Field in CI/CD Report: The ‘Overall Fail Criteria’ column is displayed now in CICD new report format where the user can see the criteria for overall failure of test. For example: The overall test is marked failed if 10% and more transactions are critical.

Pattern Option in Check Profile while Selecting Multiple Indices: An option is provided for pattern in check profile UI. The user can now provide pattern for selecting multiple indices at a time.

Time Frame Enhancements: Now four options are displayed in NS-ND plugin for selecting time period i.e. Whole scenario, absolute time, elapsed time, and phase. Earlier, the time period setup on the post-build action in Jenkins UI was in Absolute Time Period.

Allowing Decimal and Zero in Threshold Values: Now, 0 and decimal value are allowed for critical threshold value, warning threshold value, and overall fail criteria. Decimal is limited to two places only. Earlier, when the user tried to enter zero or decimal, an error message was getting displayed.

Sending TestSuite Report from NetStorm: A new tab is provided in Test Case configuration named Mail Configuration for configuring email. The user can select the ‘Send Report as Attachment’ and provide To, Cc, and Bcc fields. On report generation, an email is sent to the mentioned users in To, Cc, and Bcc fields.

Optional Baseline Time Period option in Jenkins Publisher: In ND and NS-ND plugin, the ‘Baseline Time Period’ option is optional now. If the user leave this option blank, the baseline test run column is not displayed in report.

Dashboard

Widget Wise Update: Business dashboards provide visualizations of business data to communicate metrics and other key points to users at a glance. Such a dashboard should respond quickly to any dynamic user interaction. The loading performance of a dashboard not only affects the dashboard objects’ rendering but also the connected database’s querying performance. Large dashboards can take a long time to render, especially when widgets require more graphs to be loaded. Quick rendering of data into the widgets and display in desired format is the key enabler for performance improvement.

Previously, whenever the Dashboard was opened, the default favorite configured for the user was displayed, and the favorite’s data was loaded in one shot for all the widgets/panels under the favorite. It is the case with sample update on regular interval, the whole favorite data is loaded and the whole favorite widgets/panels data is rendered.

This works fine when the loaded favorite is small, but it takes time when the favorite has many widgets/panels and that too with lots of metrics in each panel.

We have seen that in Production environment, teams are creating huge favorites and hence even favorite load takes minutes. Even if we divide the favorite in multiple pages, then also all the pages data is currently loaded whenever favorite is loaded or data update is happening, or time period is applied.

To boost performance of data loading and updates, we have changed the approach the favorite’s data is now loaded in chunks and favorite panel/widget seems to be a good fitment for the definition of a chunk.

The user is directly benefitted with the performance improvements implemented in this version:

• Data loading time in widgets, which earlier was quite large, is now minimized as individual request is sent for each widget.
• Sample update with regular time interval are much improved as compared to previous versions.
• Widgets with less data are loaded in less time as compared to the widgets having large data.
• Dragging and resizing of graphs are now improved with Design mode.

Summary: Large number of panels (>50) with each panel having large number of graphs

• Issues
  • Lot of data points need to be picked from metric DB and sent to GUI takes time
  • Large data transfer (> 30 MB) over Network takes time
  • All panels displayed together after data received – rendering takes time
• Solution – Target to show all panels in Viewport in less than 5 secs
  • Panel-wise data request and rendering
  • Lazy loading of data for panels not visible

Alert: BigPanda Integration: Alerts are now integrated with BigPanda. BigPanda is an algorithmic event management platform that automates how IT teams manage the explosion of alerts and incidents they face every day. By intelligently correlating massive volumes of alerts from fragmented clouds, applications, and infrastructure, BigPanda enables IT Operations teams to cut through the noise, remediation of IT issues faster and with fewer resources, and achieve dramatically higher service levels.

Support for ACL and OKTA login System for Machine Learning (ML) Dashboard: Now, the ML Dashboard login system is more secured with the support of OKTA. It now can be accessed by using the ND login (for both OKTA and ACL). On hitting the ML IP, the user is redirected to OKTA or ACL login page depends on the configuration at NetDiagnostics (ND) side. The user needs to provide the login credentials, which are validated, before redirecting the user to ML Dashboard application.

NetDiagnostics

Configuration UI: Dynamo Support

DB Entry Point Support: Support for below Dynamo DB entry point is provided:

com.amazonaws.services.dynamodbv2.AmazonDynamoDBClient.doInvoke(Lcom/amazonaws/Request;Lcom/amazonaws/http/HttpResponseHandler;Lcom/amazonaws/http/ExecutionContext;Ljava/net/URI;Ljava/net/URI;)Lcom/amazonaws/Response;|dynamoDB|1

Callout Support: The Dynamo callout support is provided in Transaction FlowMap, Method Call Tree, and BT IP Summary Report. The user can now get dynamo DB calls in Transaction FlowMap, MCT, and BT IP Summary.

Pagination Support for Thread Hotspot Request: This enhancement is related to performance improvement. Earlier, all the data is loaded at once, which took lots of time in rendering, but now data is loaded in pages. If the user needs data for another page, a new request is encountered to fetch the data for that page and displayed accordingly. This improves the overall performance of displayed thread hotspot data.

NetVision

Display of Page Details (by default): In NVSPA, when a user clicks on any sessions, page details is displayed (by default). It can be configured via backend that session details / page details should be displayed by specifying the value of a setting as true / false.

NetForest

Query Widget Support in Dashboard: An additional feature is being added on Dashboard widget settings, which will help the user to fetch data from NetForest (NF) using NF REST APIs and displays it on Web Dashboard without opening NF UI. The users need not to login into NF to see these visualizations and data can be displayed in various formats, such as line, table, bar, and vertical line chart.

Java 11 Support (Log Agent and Machine Agent): Support for Java 11 is provided for NetForest, which earlier was limited to Java 8 only. With this enhancement, the user can now use Log Agent and Machine Agent both with Java 8 and Java 11 with all the existing features, such as pipeline monitoring, retry, reload, token-bucket etc. This has improved the overall performance.

Auto Suggestion: To ease the search operation in query string, the users are provided with auto-suggestion feature. The purpose of this feature is to search a virtually unbounded list for related keywords and phrases, which may or may not match the precise query string. For example if the user type any alphabet like ‘t’, then it should suggest all field starting with ‘t’, such as tier, type etc. It helps users conceive and articulate queries that are more effective.

Key Points:

• Facilitates novel query reformulations
• Select from an open-ended list of terms or phrases
• Encourage exploratory search. Where appropriate, complement search suggestions with recent searches

Working of ‘Append’ with ‘Lookup’: The user can now use ‘Append’ query with ‘Lookup’. Earlier, while running a query, we were getting maximum hits as 10000, which should be max 20000 as lookup was not working. After this enhancement, we are getting max 20000 hits for our query.

New Plugin ‘nd-nfparser’ to Parse ND Logs: The nd-parse plugin is now used at NFDB end to parse the logs, which is sent by ndagent. Earlier, we were using nfagent to parse and send data to nfdb. Now, nfagent is not required if ndparse plugin is present.

Drop-down List of Existing Environments: The user can set a particular value of environment, which is selected by default when NFUI loads. Earlier, NFUI use to load that env, which is present at first position inside NetForest.yml file. Click Settings > Advanced Settings > default Env > Drop-down list.

One Request at a Time: Earlier, the ‘Search’ button is always enabled for which multiple requests have been sent to NFDB at a time. Due to this, the load on the NFDB got high. To minimize these loads, we have to disable the search button until we get the results in the NFUI side.

Add Reload for Linux Platform and creating systemd service U16: This enhancement is about implementation of reload functionality and run application as a service in ubuntu. Now, we are able to reload it using nc (netcat tool) from other machine in the same network. We can also start NF-Forwarder as a service in ubuntu/linux machine.

Bug Fixes