In this blog post, I will be talking about building a reliable data injection pipeline for Snowflake.
Snowflake is a data warehouse built for the cloud. It works across multiple clouds and combines the power of data warehousing, the flexibility of big data platforms, and the elasticity of the cloud.
Based on the Snowflake documentation, loading data is a two-step process:
- Upload (i.e. stage) one or more data files into either an internal stage (i.e. within Snowflake) or an external location.
- Use the COPY INTO command to load the contents of the staged file(s) into a Snowflake database table.
It is obvious that one step is missing: preparing data files to be loaded in Snowflake.
If steps 1-3 do not look complicated to you, let’s add more details.
Typically, developers that are tasked with loading data into any data warehouse dealing with the following issues:
- How to build a reliable injection pipeline, which loads hundreds of millions of records every day.
- How to load only recent changes (incremental replication).
- How to transform data before loading into the data warehouse.
- How to transform data after loading into the data warehouse.
- How to deal with changed metadata (table structure) in both the source and in the destination.
- How to load data from nested datasets, typically returned by the web services (in addition to loading data from the relational databases).
This is just a short list of hand-picked problems. The good news is that Snowflake is built from the ground up to help with bulk-loading data, thanks to the very robust COPY INTO command, and continues-loading using Snowpipe.
Any Snowflake injection pipeline should at least be utilizing the COPY INTO command and, possibly Snowpipe.
The simplest ETL process that loads data into the Snowflake will look like this:
- Extract data from the source and create CSV (also JSON, XML, and other formats) data files.
- Archive files using gz compression algorithm.
- Copy data files into the Amazon S3 staging bucket (also Azure blob).
- Execute COPY INTO command using a wildcard file mask to load data into the Snowflake table.
- Repeat 1-4 for multiple data sources. Injection ends here.
- If needed, execute SQL statements in Snowflake database to transform data. For example, populate dimensions from the staging tables.
The part where you need to build a “reliable data injection pipeline” typically includes:
- Performance considerations and data streaming.
- Error-handling and retries.
- Notifications on success and failure.
- Reliability when moving files to the staging area in S3 or Azure.
COPY INTO command can load data from the files archived using gz compression algorithm. So, it would make sense to archive all the data files before copying or moving them to the staging area.
- Cleaning up: what to do with all these data files after they have been loaded (or not loaded) into the Snowflake.
- Dealing with changing table structure in the source and in the destination.
Snowflake supports transforming data while loading it into a table using the COPY INTO <table> command but it will not allow you to load data with inconsistent structure.
Add the need to handle incremental updates in the source (change replication) and you got yourself a [relatively] complicated project at hands.
As always, there are two options:
- Develop home-grown ETL using a combination of scripts and in-house tools.
- Develop solution using third-party ETL tool or service.
Assuming that you are ready to choose option 2 (if not, go to paragraph one), let’s discuss
The requirements for the right ETL tool for the job
When selecting the ETL tool or service the questions you should be asking yourself are:
- How much are you willing to invest in learning?
- Do you prefer the code-first or the drag&drop approach?
- Do you need to extract data from the semi-structured and unstructured data sources (typically web services) or all your data is in the relational database?
- Are you looking for point-to-point integration between well-known data sources (for example, Salesforce->Snowflake ) with the minimum customization, or you need to build a custom integration?
- Do you need your tool to support change replication?
- How about real-time or almost real-time ETL?
- Are you looking for a hosted and managed service, running in the cloud or on-premise solution?
Why Etlworks is the best tool for loading data in Snowflake?
First, just like Snowflake, Etlworks is a cloud-first data integration service. It works perfectly well when installed on-premise, but it really shines in the cloud. When subscribing to the service, you can choose the region that is closest to your Snowflake instance which will make all the difference as far as the fast data load is concerned. Also, you won’t have to worry about managing the service.
Second, in Etlwoks you can build even the most complicated data integration flows and transformations using simple drag&drop interface. No need to learn a new language and no complicated build-test-deploy process.
Third, if you are dealing with heterogeneous data sources, web services, semi-structured or unstructured data, or transformations which go beyond the simple point-to-point, pre-baked integrations – you are probably limited to just a few tools. Etlworks is one of them.
Last but not least, if you need your tool to support a native change (incremental) replication from relational databases or web services, Etlworks can handle this as well. No programming required. And it is fast.
How it works
In Etlworks, you can choose from several highly configurable data integration flows, optimized for Snowflake:
- Extract data from databases and load in Snowflake.
- Extract data from data objects (including web services) and load in Snowflake.
- Extract data from well-known APIs (such as Google Analytics) and load in Snowflake.
- Load existing files in Snowflake.
- Execute any SQL statement or multiple SQL statements.
Behind the scene, the flows perform complicated transformations and create data files for Snowflake, archive files using gz algorithm before copying to the staging area in the cloud, automatically create and execute COPY INTO <table> command, and much more. For example, the flow can automatically create a table in Snowflake if it does not exist, or it can purge the data files in case of error (Snowflake can automatically purge the file in case of success).
You can find the actual, step-by-step instructions on how to build Snowflake data integration flows in Etlworks in our documentation.
The extra bonus is that in Etlworks you can connect to the Snowflake database, discover the schemas, tables, and columns, run SQL queries, and share queries with the team. All without ever using Snowflake SQL workbench. Even better – you can connect to all your data sources and destinations, regardless of the format and location to discover the data and the metadata. Learn more about Etlworks Explorer.
In this blog post, I will explain the various Etlworks installation options and how our pricing model works.
Our pricing model is a three-tiered, feature-based SaaS subscription.
There are three clearly defined tiers: Startup, Business and Enterprise.
All tiers allow for an unlimited number of users and provide you access to all available connectors. Tier differences can be found in the following:
- The total number of records which can be processed each day.
- The total number of schedules.
- How often the scheduled flow can run.
- The type of instance: shared or dedicated.
- The ability to set up a custom domain with its own SSL certificate
- White labeling
- The ability to create and use tenants.
Take a look at our pricing grid for more information.
Please note: Dedicated instance, custom domains, white labeling and the ability to use tenants are only available with the Enterprise plan.
Q. What are tenants?
A. Essentially, tenants are sub-accounts under the main account.
Each tenant has a separate list of users, flows, connections, formats, and listeners, and are completely isolated from each other. They can be used to separate customers and environments – for example, DEV, QA, PRODUCTION, etc. Flows and connections can be copied from one tenant to another.
Q. What happens if we exceed the maximum allowed number of processed records?
A. You will receive a notification that the limit has been exceeded and we will recommend that you switch to a higher tier.
It’s OK if you exceed the limit set for your plan once or twice a week, but systematically exceeding it violates the terms of service.
Q. How does the free trial work?
A. You get full functionality for 14 days, at which point you can turn your trial into a monthly or annual subscription.
Q. What happens when the 14-days free evaluation period ends? Do I need to ask to extend it?
A. The countdown will be reset and you will be able to continue evaluating Etlworks. We only manually disable accounts when users violate terms of service or stop using the application.
Q. How does your on-premise offering work?
A. Unlike cloud subscriptions, the customer owns and operates the on-premise instance running Etlworks. In addition to the subscription fee, we charge $2000 USD as a one-time installation and configuration fee.
Typically we don’t have access to the on-premise instances at all, but we do provide a fully automated, one-click installation and upgrade script for all major operating systems.
The on-premise instance must be able to connect to the Etlworks License server at least once a day.
Q. Can I have more than one dedicated instance?
A. Yes, you can. Each instance requires a separate license. There is a 50% discount for each additional instance you purchase.
Q. Can Etlworks handle our load?
A. Etlworks is extremely fast and optimized for performance. It is also horizontally scalable. You can have multiple ETL servers running in parallel behind the load balancer. There is a 50% discount for each additional ETL server you purchase.
Q. What is ETL server?
A. The ETL server is a headless web application with the REST API which handles the data integration requests. It is a core of the Etlworks Integrator. The ETL server can run on the same physical instance as Etlworks Integrator (this is a default setup), or it can be installed on the multiple physical/virtual servers behind the load balancer, which will create a highly scalable data integration solution that is able to process a large number of heavy-duty ETL requests.
Q. How many ETL servers will I need?
A. In most cases, you will need just one ETL server, which is included in the base price. You might need more than one ETL server if you expect a very large number of the parallel ETL requests (hundreds of thousands per day). Always think of upgrading the amount of available RAM before adding an extra ETL server.
Q. Is multi-server option available for cloud Enterprise plans?
A. Yes, it is available.
Q. What does “the price starts from” mean for cloud Enterprise plans?
A. We factor the cost of running the dedicated instance of Etlworks when calculating the price. For example:
- 8 Gb RAM, 2 CPU cores, 100 SSD – $750 / month.
- 16 Gb RAM, 4 CPU cores, 100 SSD – $950 / month.
- 32 Gb RAM, 8 CPU cores, 100 SSD – $1500 / month.
Q. How can I estimate the size of the instance?
A. If you are planning to work with APIs that return large JSON or XML documents we recommend the instance with 16 or 32 Gb of RAM. In all other cases, 8 Gb should be enough.
Q. Can I request to upgrade or downgrade the instance?
A. Yes, you can. It usually takes just a few minutes to change the size of the instance for the instances managed by Etlworks.
Q. What does “the price starts from” mean for on-premise Enterprise plans?
A. The base price includes one instance of Etlworks Integrator. Each additional instance or each additional ETL server is provided with 50% discount. For example:
- One instance with one ETL server – $750 / month.
- Two instances (for example DEV and PRODUCTION), each with its own ETL server – $1200 / month.
- One instance with two ETL servers – $1200 / month.
- One instance with three ETL servers – $1600 / month.
Q. Can I download and install Etlworks from the AWS or Azure marketplace?
A. Not at the moment. We will be in both marketplaces by the end of 2018.
Q. How am I billed?
Q. What will my total cost be?
A. Your total cost will be the subscription cost listed on our website. No surprises.
Please note: Enterprise plans will be billed based on the size of the instance.
Q. What regions are available for shared and dedicated instances?
A. We have shared instances in the US EAST (us-east-2) and EU WEST (eu-west-2) regions. Dedicated instances can be installed in any region of your choice.
Q. On your website there is an option to buy a perpetual license. What does this mean?
A. When you buy a perpetual license you own it forever and don’t need to pay for the subscription. Perpetual license for one server costs 30K.
There is an option to subscribe to upgrades after the first year- they are free during the first year.
Q. How is support provided?
A. The basic support by email is included in all plans. You can purchase an extended support, which includes 10 hours of the professional services per month for $600 / month. You can also buy ad-hoc extended support for $150 / hour in 4-hours batches: 4, 8, 12, 16, etc.
In this blog post, I will discuss the definition of cloud data integration and what makes it truly useful.
Before we start, let’s get on the same page and define what cloud data integration is.
According to Wikipedia, cloud data integration software must have the following features:
- Deployed on a multi-tenant, elastic cloud infrastructure.
- Subscription model pricing: operating expense, not capital expenditure.
- No software development: required connectors should already be available.
- Users do not perform deployment or manage the platform itself.
- Presence of integration management & monitoring features.
While I agree with the definition, there’s something is missing:
where is the data we are suppose to be integrating?
If you are ahead of the curve, all your data is already stored in the cloud. While I think we all will be here eventually, as of today, a typical enterprise – from two guys working out of a garage to multinational corporations – owns and operates multiple data silos. I would add diverse and isolated data silos:
- Cloud databases.
- On-premise databases, available from the Internet.
- On-premise databases, not available from the Internet.
- Public APIs.
- Private APIs, not available from the Internet.
- Cloud-based third-party applications.
- Locally hosted third-party applications.
- Legacy applications.
- Files stored locally.
- Files stored in cloud data storage.
- Office 365 and Google Docs documents.
Can your favorite data integration platform handle the vast array of data sources? If the answer is “Yes it can! We just need to deploy it in our corporate network and it will be able to connect to all our data,” then it is not a cloud data integration anymore. Don’t get me wrong, there is nothing wrong with the ETL tool deployed locally. It gets the job done, but you are not getting the benefits of the true cloud-based platform, specifically this one:
users do not perform deployment or manage the platform itself.
If this is not a showstopper, my advice is to find and stick to the tool which has all the required connectors and is easy to program and operate. Sure, you will need a competent DevOps group on payroll (in addition to the ETL developers), who will be managing and monitoring the tool, installing upgrades, performing maintenance, etc., but hey…it works.
Keep reading if you want to focus on breaking the data silos in your organization instead of managing the data integration platform. The solution to the problem at hand is so-called hybrid data integration.
Hybrid data integration is a solution when some of the connectors can run on-premise, behind the corporate firewall, while others, and the platform itself runs on the cloud.
We, at Etlworks, believe that no data silo should be left behind, so in addition to our best in class cloud data integration service we offer fully autonomous, zero-maintenance data integration agents which can be installed on-premise, behind the corporate firewall. Data integration agents are essentially connectors installed locally and seamlessly integrated with a cloud-based Etlworks service.
Let’s consider these typical data integration scenarios:
Source and destination are in the cloud
Example: the source is an SQL Server database in Amazon RDS and the destination is a Snowflake data warehouse.
In this case, no extra software is required. Etlworks can connect to the majority of the cloud-based databases and APIs directly. Check out available connectors.
The source is on-premise, behind the firewall and the destination is in the cloud
Example: the source is locally hosted PostgreSQL database, not available from the Internet, and the destination is Amazon Redshift.
In this scenario, you will need to install a data integration agent as a Windows or Linux service in any available computer in your corporate network (you can install multiple agents in multiple networks if needed). The agent includes a built-in scheduler so it can run periodical extracts from your on-premise database and push changes either directly to the cloud data warehouse or to the cloud-based data storage ( Amazon S3, for example). You can then configure a flow in Etlworks, which will be pulling data from the cloud data storage and loading into the cloud-based data warehouse. The flow can use the extremely fast direct data upload into the Redshift available as a task in Etlworks.
The source is in the cloud and the destination is on-premise, behind the firewall
Example: the source is a proprietary API, available from the Internet and the destination is a database in the Azure cloud.
Data Integration Agent can work both ways: extracting data from the sources behind the firewall and loading data into the local databases. In this scenario, the flow in Etlworks will be extracting data from the cloud-based API, then transforming and storing it in the cloud-based storage, available to the agent. The data integration agent will be loading data from the cloud-based storage directly into the local database.
The source is in the cloud and a destination is a third-party software hosted locally
Example: the source is a proprietary API, available from the Internet and the destination is locally hosted data analytics software.
If the third-party software can load data from services such as Google Sheets, you can configure a flow in Etlworks, which will be extracting and transforming data from the API and loading into the Google Sheets. The third-party software will then be loading data directly from the specific worksheet. You can always find a format which is understood by Etlworks and a third-party software.
Source and destination are on-premise, behind the firewall
Example: the source is a proprietary API, not available from the Internet and the destination is another proprietary API, again not available from the Internet.
In this case, you probably don’t need cloud-based software at all. Look at Etlworks on-premise subscription options as well as an option to buy a perpetual license.
If you are reading this, you have most likely already answered the question:
“why do I need a data integration tool anyway?”
The answer is probably any combination of the things below:
- I’m tired of dealing with the complicated inputs and outputs and don’t want to write a custom code anymore.
- I need a way to connect to all my data, regardless of its format and location.
- I need a simple way to transform my data from one format to another.
- I need to track changes in my transactional database and push them to my data warehouse.
- I need to connect to the external and internal APIs with different authentication schemas, requests, and responses.
- I want to create new APIs with just a few mouse clicks, without writing any code.
- Generally speaking, I’d prefer not to write any code at all.
- I don’t want to worry about backups, performance, and job monitoring. I want someone to manage my data integration tool for me.
If this sounds familiar, you are probably in the process of selecting the best tool for the task at hands. The truth is, most modern data integration tools are quite good and can significantly simplify your life.
Etlworks is so much better.
By selecting Etlworks as your data integration tool, you will be able to implement very complex data integration flows with fewer steps and faster.
The key advantages of the Etlworks:
- It automatically parses even the most complicated JSON and XML documents (as well as other formats) and can connect to all your APIs and databases.
- It is built for Cloud but works equally well when installed on-premise.
- You can visualize and explore all your data, regardless of the format and location, before creating any data integration flows.
- You can use SQL to extract and flatten data from nested documents.
- Etlworks is a full-blown enterprise service bus (ESB) so you can create data integration APIs with just a few mouse clicks.
- Etlworks can integrate data behind the corporate firewall when working together with data integration agent. Read more in this blog post.
- We provide world-class support for all customers.
- Our service is so affordable that you won’t have to get a board approval in order to use it.
- No sales call necessary! Just sign up and start using it right away.
Test drive Etlworks for 14 days free of charge.