Share Sisense web application externally (Windows only)
Sisense dashboards can be shared easily within the same private network, but the Sisense web application is not configured to be publicly accessible by default. This page details the steps and best practices to make your Sisense instance accessible to anyone on the internet, so that dashboards can be shared with users outside of your organization and Sisense can be accessed by users not on the same local network as the Sisense server. * Note: These settings make your Sisense web application publicly accessible. Users will still need to authenticate and be shared on a dashboard to access Sisense content. Implementation Steps To access Sisense using the Sisense server's IP address (for example, "192.0.2.0"), confirm that the server has a static public IP address. Consult your IT professional, internet service provider (ISP), or hosting service for more details. To access Sisense using a subdomain/domain name (for example, "analytics.example.com"), update your DNS to point to the machine's IP. You must own and control the domain name you wish to use. Sisense does not manage or assist with domain names or their configuration. Open the inbound TCP port in your firewall. This port is 8081 by default but configurable in the installation process (https://docs.sisense.com/win/SisenseWin/installing-sisense.htm#CustomizingtheInstallation). Nominally, this can be controlled in Windows Firewall with Advanced Security, but many customers have superseding rules from their organization or hosting service. If Sisense is still not accessible over your port after opening it in Windows Firewall, confirm with your IT department whether firewall rules may be in place elsewhere. Implement a security layer using SSL to ensure that all network traffic between your Sisense server and client machines is encrypted. For steps to do so, see here: https://docs.sisense.com/win/SisenseWin/setting-up-ssl-for-sisense-windows.htm. After implementing SSL, ensure the port used (typically 443) is also open for inbound TCP connections. Once these settings are implemented, external users will be able to access Sisense by entering the Sisense DNS/IP address and port (for example, "192.0.2.0:8081" or "analytics.example.com:8081"). * Note: If you don't want users to have to enter a port (for example, "192.0.2.0" or "analytics.example"), use port 80 for http communication and 443 for https communication, as these ports will be hidden from the URL. Troubleshooting If a user is unable to connect remotely after the settings were applied, you can test the communication between the machines to understand the source of the connectivity failure by doing the following: Confirm the Sisense web application is running and accessible locally on the server Use the 'nslookup' command in the Windows Command Prompt (cmd.exe) to confirm your domain/subdomain points to the correct IP address for your Sisense server. For example, 'nslookup example.com'. Try pinging the machine's IP address in the command prompt. This will show if there is communication available between both machines. For example, 'ping 192.0.2.0'. If you are able to ping the machine, use the telnet command to ensure you are able to communicate through the required port. For example, 'telnet 192.0.2.0 8081' Further troubleshooting or review may require your IT team's assistance. Further Reading More information is available here: https://docs.sisense.com/win/SisenseWin/setting-up-your-production-environment.htm. Helpful general information about web applications and security is available from Mozilla: https://developer.mozilla.org/en-US/Apps https://developer.mozilla.org/en-US/docs/Web/SecurityHow to Export a Certificate using the Certificate Export Wizard (Windows only)
How to Export a Certificate using the Certificate Export Wizard In Sisense version 7.2 and later, SSL (Secure Sockets Layer) settings have moved from Microsoft IIS into the Sisense Configuration Manager. For organizations moving to Sisense 7.2 and above from an earlier version that would like to continue using SSL, it is necessary to obtain the certificate to make SSL work. See the steps below to attempt to obtain the certificate from the server's certificate store if your organization previously had SSL set up on the Sisense web server. Organizations can also opt to contact their Certificate Authority (CA) to obtain an up-to-date copy of your site's .pfx or .crt and .key files. This explanation utilizes mmc (Microsoft Management Console) to export the certificate. On your webserver, search programs for mmc and select mmc.exe In the window that appears, click on "Add/Remove Snap-in..." In the "Available snap-ins" column, click on Certificates then click "Add >" to add it to the "Selected snap-ins" column In the resulting pop-up, select "Computer account" then click 'Next' Leave the settings as is on the next screen and click Finish Click on OK to add the Certificate snap-in On the left-hand side of the screen, expand Certificate (Local Computer) > Personal > Certificates Right-click on your certificate, go to All Tasks > Export The Certificate Export Wizard will appear which will assist you in exporting your organization's certificate to the appropriate format. Select the Next button. On the Export Private Key screen, select "Yes, export the private key". The private key is necessary for SSL to work in Sisense 7.2 and higher. Then select the Next button. On the Export File Format screen, select the Personal Information Exchange (.PFX) file option and hit the check boxes that correspond to the options below, then select the Next button. ☐ Include all certificates in the certification path if possible ☐ Export all extended properties ☐ Enable certificate privacy On the Security screen, select the Password checkbox and enter a password. Make sure the encryption setting is TripleDES-SHA1. Then select the Next button. Select the file location in which to save the exported certificate. Then select the Next Button. On the last screen select the Finish button to complete the setup. Complete the setup in this guide to set up SSL in Sisense.Host a New Webpage on IIS (Windows only)
It some situations, it may be useful to be able to use IIS to host website that work with Sisense. A common example would be hosting a webpage that contains the code for an embedded dashboard. Additionally, it could be useful in scenarios when you want to host a web page to be used as your custom home page during rebranding, or when hosting a webpage for the iframe widget This article shows a simple way to host your own webpages using IIS. 1. Add a New Directory to IIS The directory we will add to IIS needs to exist like any other file path, and needs to contain the webpage we want to use. We will create a reference to this directory in IIS so it can be accessed by other users, and parsed correctly by a browser. To do this, open IIS, right click on SiSenseWeb, and go to Add Virtual Directory For this example, I created a new folder to hold all of my webpages at C:\embedded, and I gave it the alias of 'testpages' in IIS. Since I used the default installation settings (localhost:8081), the file C:\embedded\adventure_works.html will be accessible at localhost:8081/testpages/embedded/adventure_works.html. After this, right click the new virtual directory you've created, go to Edit Permissions..., go to the Sharing tab, and click share. From here, use the drop down list to find Everyone, and give them Read access permissions. If a popup appears asking if you like to turn on network discovery, click Yes. In this example, I'm going to host a simple html web page that uses an iframe to display an embedded dashboard. This webpage can be created with notepad. Just be sure to save it with a .html file extension. If you are unfamiliar with embedding dashboards, you can find more information about them here: https://docs.sisense.com/main/SisenseLinux/embedding-dashboards-and-widgets.htm In the example from the documentation, the basic iframe code to embed a dashboard will look something like this: <iframe id=”ifm” name=”ifm” width=”100%” height=”100%” frameborder=”0″ src=“http://mysite.com/app/main#/dashboards/536f3a54d093e26c28000114/widgets/536f3a54d093e26c2800011b?embed=true″ scrolling=”auto”></iframe> However, you'll need to wrap this code in some standard html tags, so it gets parsed correctly by the browser. These tags will give you the format of: <html> <head> </head> <body> <iframe id=”ifm” name=”ifm” width=”100%” height=”100%” frameborder=”0″ src=“http://mysite.com/app/main#/dashboards/536f3a54d093e26c28000114/widgets/536f3a54d093e26c2800011b?embed=true″ scrolling=”auto”></iframe> </body> <html> Now, save the this file in the physical directory you created to be used by IIS. (In my case its C:\embedded). Now we can access this embedded dashboard, with the html code that's attached to this post.How To Find Your Sisense Version? (Windows only)
Question: How can I find the exact version and build of my Sisense environment? Solution: There are 3 easy ways to find the exact version of your environment: On the web application Open your Sisense web application On the top toolbar, click on the person icon. The version will be the last item on the menu which pops up: On the Desktop ECM Note: Not relevant for Sisense Linux Open the desktop Elasticube Manager Click the Help & How-To's button on the top scroll bar Click the About button The version will be here: On the web application Open dev tools - console (F12) Copy this and press enter: prism.version The version will be displayed as output:Resizing System Volume Information (SVI) Windows only
Question: Sometimes your server might run out of space on because the SVI (System Volume Information) is eating up all memory. The system volume information is used in windows in order to save information related to system restore which is used when a restoration of the system is performed. The problem is that if there is no size limit to this feature, it might take a significant amount of the system memory, without ability to directly delete it. Answer: There are two ways to resize the SVI: Through the system properties - follow these instructions: http://www.pcadvisor.co.uk/how-to/windows/how-make-system-volume-information-folder-smaller-3613911/ Through the command line - if you don't see the system protection tab according to the above method, you can resize it through the command line: https://indrajitc.wordpress.com/2008/03/25/reclaiming-disk-space-from-system-volume-information/REST API - Adding Data Level Security
When building code to automate the process of adding users (or groups), it may be beneficial to add security around those users. Follow the steps below to learn how to add data level security through the REST API: Step 1 From your SiSense home page, navigate to the Manage tab and choose the option for REST API. From here, click on the link to the REST API Reference interface. Step 2 From here, choose the 0.9 version and expand the Elasticube section and scroll down to the POST /elasticubes/datasecurity section. Step 3 The sample code below shows a valid JSON object to use as part of the REST API request. Starting from the sample code, replace the value for party with a user identifier that already exists in your system. Then paste the code into the REST API interface and click run. You should see a response of 200, which indicates a successful operation. [{ "server": "LocalHost", "elasticube": "Sample Lead Generation", "table": "Lead Generation", "column": "Country", "datatype": "text", "shares": [{ "party": "5c0f85690ca2f66cc242e266", "type": "user" }], "members": [ "United States", "England" ] }] Notes The server name value is case sensitive, so make sure it matches with your system. The sample provided is specific to one of the sample databases that comes with SiSense. To modify this to a different system, just update necessary fields to include your specifications. The sample provided is an array that consists of a single element. This could be modified to contain several entries all sent at once.Custom UI Colours (Instead Of Yellow)
Download: For V7.0 - V7.1, For V7.2 and V7.3, For V7.4 Plugin allows changing Sisense yellow color to any color you wish. This plugin causes instability with some Sisense versions and is not supported. Use at your own risk. STEPS STEP 1 - ADD THE PLUGIN Download the attachment and unzip the contents into your C:\Program Files\Sisense\PrismWeb\plugins\ folder. If the plugins folder doesn't exist, just create it. After those files have been unzipped there, you may also have to restart the web server. STEP 2 - USE THE PLUGIN Reload sisense web app. BEFORE: AFTER: CONFIGURATION: Navigate to .../styles in the plugin folder and edit config.less (or config.scss) file with colors according to your desire. Change log: 05/17/2019 - Support for V7.4 and change color for button on login page. NOTE: In the latest versions of Sisense (Windows and Linux - 8.2+), you have this feature baked into the product - Look & Feel (on the Admin tab)Quickly Add Unique Identifier to Blox Action Buttons Without Modifying Blox Template
A customer recently had the unusual request for a method to select and distinguish, via a CSS selector used within a custom Javascript action, for a specific Blox buttons in a existing Blox widget that contained multiple identical buttons, identical in inner text and all other parameters, without adding any new lines to the Blox template to add unique identifiers.Personio API Pre-Processing Data
Question Is there a way to add pre-processing (python script) to the data that will be uploaded from a custom API connector? Or should I run the script locally on a fixed schedule and use the Sisense API to upload after the data is processed? Answer Some info about your questions (as we already come across with that many times before): You can run any python script that you would like automatically before or after a Build is done using an ElastiCube Plugin , scroll to the part where is talks about python scripts. Sisense doesn't provide an API to modify data within your cube. So if you would like the Python script to do some extra processing before it gets pushed to the cube - you can save that data locally on the Sisense machine in csv format and then use another another post-build plugin to delete that csv once the build is finished. - Notice that you can use any python version you'd like (in their code sample they are using the old 2.7)Maximize Dashboard Performance
What is good dashboard performance? Dashboard performance is the time it takes to load a Sisense dashboard. We want our users to have a good experience when working with our data. The time taken to load the dashboard is the first step. Each individual widget in a dashboard is a separate query. The dashboard needs to send these queries through the Microsoft IIS web server to the ElastiCube, pull back the results and populate the data visualizations. The questions we will address in this post are what is a reasonable dashboard load time and how can we enhance performance. So, what is good dashboard performance? This is a slightly subjective question. Users always want the fastest load time. However, we must think about load time relative to the data we are processing. If the dashboard is aggregating billions of rows of data, it is unrealistic to expect two-second load times. The converse is true as well — if a dashboard is based on a few thousand rows of data, the dashboard should not take five minutes to load. The goal, to put it simply, is to optimize the load time relative to the amount of data the dashboard is processing. Sisense’s architecture is designed to return queries very quickly, and there are ways to help it. How to monitor performance Timing First and foremost, how long does the dashboard take to load after the user navigates to the URL? Does the load time inhibit the user experience? As discussed before, the answer will vary by the use case, ElastiCube size, and end user. However, as a general guideline we want the load time to be under a minute if the dashboard is not crunching enormous amounts of data. System Resources Monitoring system resources during dashboard load is the first step to identifying possible issues. These can be monitored live by opening the Task Manager > Performance tab. RAM and CPU should not be maxed out during the dashboard load. If they are, it could be hindering the process. More system resources might be necessary (see here for Sisense's minimum hardware guidelines). Components ElastiCube.exe – if this process is taking CPU and RAM then the query is too heavy. This could be a possible many to many occurrence (covered below) ElastiCubeManagement Service – if this process is taking CPU and RAM, you most likely have a build occurring at the same time Java(TM) Platform SE Binary - This process transmits Sisense monitoring data. Heavy use of hardware is expected when restarting the service after a while (as it catches up on transmitting data), but if you experience continued consumption, confirm you are using the latest version of the Sisense Monitor Node.js - If this process is heavily consuming resources, check whether you have recently published dashboards to a large number of users. w3wp / iisexpress – if this process is taking CPU and RAM you may be overloading your IIS server. This is mostly likely when presenting too much data (covered below) Another way of monitoring system resources is by using the Windows Performance monitor. This will help analyze system resources over time. The post here should help: https://learn.microsoft.com/en-us/windows-server/identity/ad-fs/deployment/configure-performance-monitoring Timing vs other dashboards Using other dashboards that use similar data sets as a baseline can also be useful. For example, two dashboards are using the same ElastiCube. Dashboard-A takes 30 seconds to load. Dashboard-B takes five minutes. Knowing that Dashboard-A loads normally, we can tell that Dashboard B has some kind of an issue. Possible issues could be a bad join, many-to-many relationship, or specific troublesome widget that Dashboard A did not have. Compare the differences to narrow down possible issues. Web Developer Console Developer Tools within your web browser allow us to track each dashboard component's load time. In Chrome before loading the dashboard, open the developer console (F12 in Chrome). Navigate to the Network tab. From there you can see the timeline and the long-running requests. Most likely it will be the JAQL queries that take the longest. You can then click those records for more information. Under Headers>Request Payload there is information about the widget ID and Name. This will help identify the long-running queries. Under Timing, you can view what composed the query response time. Stalled represents one query waiting for other queries to finish (there is a limit to the number of concurrent queries your web browser can run). Waiting (TTFB) represents time waiting for the ElastiCube to respond. Content Download represents the amount of time retrieving the query results. If the JAQL queries are not the longest running components, you may have connectivity issues to your server, or are using too much Javascript (covered below). Causes of slowness System Resources A reasonable query time depends on a few factors. First and foremost, system resources. The recommended system resources for Sisense can be found here: https://docs.sisense.com/win/SisenseWin/windows-minimum-requirements.htm. Amount of data The amount of data crunched/returned in a query has an impact on performance. Logically, this makes sense. Crunching one million rows will, of course, return faster than one billion rows. The way to affect this is by adjusting filters to limit the data returned. Furthermore, limiting the amount of data in the ElastiCube is useful. This can be done by editing either the tables that Sisense is connected to, or by editing the SQL statement that is used to query the original data source. Schema Design Schema design is a very important factor in query performance. Each data connection within the ElastiCube will become an inner join when the dashboard queries this data. Joins are costly when it comes to query time. When designing the schema for performance, the ElastiCube designer should de-normalize the schema as much as possible. De-normalization means having a fewer amount of joins, and including redundant information in one table. For more on schema de-normalization, see here. Custom Calculations in the Dashboard Custom calculations are often useful when manipulating data to answer business questions. They allow users to aggregate data in various ways to present visually. However, they do take time to process. Each custom calculation is a separate query that is processed by Sisense. To enhance query performance we can cut down the number of queries generated by the dashboard. This is done by moving some calculations to the ElastiCube, thus moving that process time to the build time instead of dashboard load. The best example is a row calculation, such as the addition of two separate columns — FieldA + Field B. This type of calculation can be easily added to a table using the same exact formula as a custom field. If it is from a different table, we can use the LOOKUP function in the ElastiCube Manager. Presenting too much data Presenting large amounts of data is costly on dashboard load time as well. More specifically this causes slowness when using a pivot or table widget with many rows. The web server has to process the large amount of data and store it to allow the tool to present it visually. If it is possible, limit the number of dimensions that break the data to such a granular form. Sisense generally recommends 6-8 widgets per dashboard for optimal performance. Each widget incurs incremental query and rendering time. Sisense's official add-ons such as Accordion, Jump to Dashboard, and Switchable Dimensions may be leveraged to reduce the number of widgets while preserving the amount of data accessible to end users. When an end user loads the first dashboard in a session, performance will be slower against subsequent dashboard loads, as common web application elements such as logos, toolbars, and code templates are loaded for the first time. Later navigation between dashboards once these resources are cached will be faster. Other causes of query slowness and slow dashboard performance Many to Many If RAM grows exponentially, and the dashboard takes a very long time to load, you could have a possible many to many relationship. A many to many relationship (M2M) can cause extreme slowness in dashboard load time. A M2M is a relationship in the ElastiCube where neither side of the relationship is unique. This creates a Cartesian product, thus consuming all of the machine's resources. Relationship Cycle A relationship cycle exists when there are multiple join paths between two tables in a data model. Sisense will generally opt for the shortest path, but depending on the dimensions used (in widgets or filters), Sisense can take join paths through tables that are unexpected to dashboard designers. While not as deleterious to performance, an additional join on a large table can reduce performance and possibly lead to unexpected results. For more information on relationship cycles in data modeling, please visit the Handling Relationship Cycles documentation page. For the detection of data model issues such as many-to-many relationships and cycles, we recommend using the official Visualize Queries plugin. Data Security If some users are seeing slow performance but others are not, it is worthwhile to check the data security of the user in question. Some questions to consider: Are there complex queries occurring on the ElastiCube due to competing data security settings? Does the data brought in by data security cause pivot tables or other widgets to display a larger amount of data than expected? Please reference the official documentation on data security here. Browser Compatibility If some users are seeing slower dashboard performance on certain machines or browsers, check to make sure they are using a support browser outlined in the Minimum Requirements and Supported Platforms document.
