How to use the Desktop Cloud Automation Service
All Analytics Canvas users have access to the Desktop Cloud Automation service to schedule workflows. As a Cloud Hosted Service, it is subject to quota, which is shared with other Analytics Canvas products, including the Looker Studio Partner Data Connector and Analytics Canvas Online.
Overview of Desktop Cloud Automation
Desktop-Cloud automation is a cost effective way to automate workflows created with Analytics Canvas Desktop. Any files created with the Desktop application can be automated, provided all data sources can be accessed by the Analytics Canvas Cloud.
Generally if you are using APIs the process is straight forward. If you use reference files, such as an Excel Template, it will have to be uploaded to the cloud. If you access a database, you will need to permit our cloud to read from, or write to, your database.
Preparing your workflows for Cloud Automation
Cloud Automation means that you will upload your Canvas file to the cloud where it will run on our Analytics Canvas Cloud Servers. The servers are not synced with or accessible by your local machine - you are placing your files onto the servers where they will run.
As such, your local directory is not accessible, nor can data be written back to your desktop directly. Instead, we must connect to your data sources through remote channels, such as APIs to file sharing services, and to remotely available databases.
- Inputs from databases must use a username and password as NT Login will not work remotely from outside of your firewall.
- Canvas Cloud must have permission to access your databases. For that you must permit our IPs. Contact support@analyticscanvas.com to get the list of IPs.
- If an input file is not connected by API, it will remain static and will only change when the package is manually synced.
- Output files, such as Text, CSV, Excel,Tableau Data Extracts (TDE), or Tableau Hyper files, must be connected to one of the 4 APIs for file delivery: Amazon S3, Google Cloud Storage, Dropbox, or Google Drive. This way the file can be delivered from our cloud to you.
- Tableau users have the option of having their files uploaded directly to Tableau Online or Tableau Server.
Creating a Canvas Cloud Automation Package
To have canvases run automatically in the cloud, you create a Canvas Automation Package. This package contains one or more canvases (represented by .ACC files) that are run in order.
It is recommended that you put all the .ACC files that you want to run together, and any data files they need into a single folder on your machine.
To begin, click the Cloud Automation button on the main Canvas toolbar, or click Automation > Cloud Automation from the main menu.
If you haven't already, you will be asked to login with your Analytics Canvas Cloud account.
Click on "New Package"
Then click on the "Add Canvas" button to select which canvas or canvases you want to run. You can pick multiple .ACC canvas files at a time.
You can then reorder the selected canvases by dragging them, remove any unwanted canvases by clicking on the "X" button to the left, or add more canvases as needed using the "Add Canvas" button. Once you have the Canvases you want, in the order you want, press Next.
After this, you will need to provide any credentials for databases, and you will be shown which credentials are being used for API sources, and be able to change them if needed. Remember that for Google Analytics, or Google Docs, all the canvases in the package will use this credential, so it must have access to everything that the canvases need.
If you are using data sources such as Google Analytics, Google Docs, Dropbox, Google Drive, or destinations like Data Studio or BigQuery then Canvas will handle all the credentials for you. If you use databases or other sources that require a login, then you may be prompted for this login at this time so that the canvas can be run on a schedule.
To save the passwords locally so you do not have to reenter them, you can select "generate password file" this generates and obfuscated password file in the folder with the .ACC file. Keep this file safe as it contains your credentials.
Press next, and you can then define any schedule you want the package to run on. A schedule is not required, a package can also be run on demand.
You can also have notification emails sent when packages are run. More importantly, it is possible to send an email when packages fail- it is recommended that this option be used- it is best to receive an email only when there is a problem.
On the error handling an logging tab, you have options as to how you want any failed canvases to be handled.
If the canvases in the package rely on the previous canvases running, select "stop batch file immediately if a canvas fails" - this is the default.
However if you are running a number of canvases that are not related, you can select "run all canvases even if canvases fail".
Once these settings are as desired, press update, and the package will be published up to the server. You will see that for each canvas there is a green "In Sync" indicator. This advises you that there have not been any changes to the local canvas. You can change the local canvas file as you wish- the file in the cloud will not be updated until you sync the file.
If you want to run your package at a time other than the scheduled time, just click the "Run Now" button.
If you run a package just before it is scheduled to run, and it is still running when the schedule time arrives, then the scheduler will skip the scheduled run, and log it as "Skipped". The same package cannot run twice at the same time- once a package is running, all users will see that it is running, and the run now button will be disabled.
Updating a Canvas Package
To update or change what a canvas package does, first modify the Canvases as needed locally and test them locally, possibly using a different output file or database (having a draft version).
When you edit a canvas on your machine that has been included in a package and is running on the cloud, canvas will show you that the version of the Canvas running in the cloud is older than the version on the local machine- click on the Cloud Automation toolbar or automation menu selection and select the Canvas Package. You will see to the right the list of Canvases in the package, and their sync status.
Clicking on the Analytics Canvas icon to the far left will open the folder where the .ACC file is, clicking on the link with the file name will open a new instance of canvas and load that file.
Clicking on the Sync icon will sync that file and related credentials.
Once you have made the changes you wish to make to the Canvas, you can press either the "Sync Files" button or the individual sync button beside the canvas.
If you choose the individual button and hover over the "Publish Local Version" you will see the following. To update the cloud, press "Publish Local Version".
The other option is to undo your local changes, and copy the older version of the .acc file from the cloud to local (along with any older versions of static data files used).
After this, you go into the standard sync process- because a package potentially has more than one canvas and those canvases might be related, you are shown all the canvases and their status. It is possible to use the check box to not sync a local change. If a local change is not synced then the cloud version of that file stays the same.
Downloading Canvases from the Cloud
It is possible to download canvases and entire canvas packages from the cloud to your local machine.
Some examples of why you might want to do this:
You use more than one machine and you created the canvas and published it up to the cloud on a different machine.
You want to look at a canvas that someone else made and has shared with you.
You want to get the older version of the canvas that is running in the cloud and compare it with some other version.
When you go to the Canvas automation view (click on the Cloud Automation tool bar), if you do not have a local copy of the canvas you will see the following, with any missing canvases marked as "Not Linked". To get a copy of all the ACC files, press "Download All".
You will then be prompted for a folder- pick or create a folder to hold all the Canvas Files. It is recommended to have a separate folder for each Canvas Automation Package, rather than put all your Canvases in one folder, or have them spread out across multiple folders.
Downloading the cloud version without replacing your local version
You may want to keep or review what is in the cloud but also keep any changes you have done to your local copy of a Canvas. You can do this by downloading just the canvases you are interested in, and putting them in a different folder.
To do this, use the download button beside the specific file.
Reviewing Automation Activity
Once you have created Canvas Automation Packages and set schedules, you can at any time log in and review which packages have run, and if there were any issues.
Login to your cloud account and click on the "Cloud Automation" toolbar or menu item under the Automation menu.
Select the "Cloud Automation Runs" tab at the top of the dialog;
This shows you what Canvas packages have run and the resulting status. Check the status types you want to see, then press the "Refresh" button- for example, to see only failed jobs, uncheck the other statuses and press refresh.
To View the detailed log of a given package run, click on the "log" link to the right of the package run description.
The detailed log file will provide a summary of key activities in the running of the Canvas and any errors that were encountered.
To jump to the details about a given package click the "package" link.
Support + Troubleshooting
If you run into any issues with cloud automation, contact support@analyticscanvas.com. Be sure to include the contents of any log files if your issue relates to an automation task.