Highlights of Azure CosmosDB emulator
Cost: free of use
How is delivered: Windows Service or Docker container, exposing web endpoint on a local port
Top 3 features:
- #1 Docker support
- #2 A web explorer similar with the one available inside Azure Portal for local documents
- #3 The option to delete all the data with only one click
Pain points:
- #1 No consistency level supported
- #2 Limited to 25 fixed containers or 5 unlimited containers
- #3 Lack of support for other APIs
Download URL:
- MSI: https://docs.microsoft.com/en-us/azure/cosmos-db/local-emulator-release-notes
- Docker: https://github.com/Azure/azure-cosmos-db-emulator-docker
Credits: Azure CosmosDB Team
When I first discovered the emulator support for Azure CosmosDB I was super excited. Azure CosmosDB is not cheap and for development teams is always a challenge when they need to share a data repository.
Usually, each person in the team is working on different areas of the project and needs a different data setup. By having an emulator the team has the possibility to focus on their own area and test it locally without deploying their own instance of Azure Cosmos DB database or container. Not only the development is smooth, but also the running cost of the development environment is lower.
As you can see above, the interface is clean and provides the things that you need in the first moment when you start the emulator - the connection string to connect to it. The Explorer is similar to the one that we have inside Azure Portal, allowing us to do any type of operations inside our containers. We even have the ability to define functions, triggers and stored procedures.
The support of writing queries directly inside the browser it's nice and I used with success to understand some concepts related to how to write queries. You can save the queries defined in the emulator, but we aware that they are saved inside Azure CosmosDB containers. This will generate an additional cost (~$0.77/day) but can be used as a common query repository by the team.
There is an option that allows us to reset all the data stored inside the emulator. Useful during development phases. There is also the ability to upload content from JSON, for cases when you need to prepare your local emulator with fresh data.
The speed and the performance of the emulator are not high, but it is more than enough for development and to play around. Locally I don't see an issue with lack of support of consistency level especially because you have only one instance - no replica.
To be able to activate emulators for Gremlin API, Cassandra API, Table API or MongoDB emulator you need to start the emulator from the command line and specify "/EmulatorMongoDBEndpoint" as the parameter for MongoDB. A similar thing you need to do if you want the emulator for the other APIs. Except for the SQL API that is fully supported, the rest of the APIs are partially supported.
The Docker support for the emulator is awesome and enable you to use the emulator in different and configurations.
When I first discovered the emulator support for Azure CosmosDB I was super excited. Azure CosmosDB is not cheap and for development teams is always a challenge when they need to share a data repository.
Usually, each person in the team is working on different areas of the project and needs a different data setup. By having an emulator the team has the possibility to focus on their own area and test it locally without deploying their own instance of Azure Cosmos DB database or container. Not only the development is smooth, but also the running cost of the development environment is lower.
As you can see above, the interface is clean and provides the things that you need in the first moment when you start the emulator - the connection string to connect to it. The Explorer is similar to the one that we have inside Azure Portal, allowing us to do any type of operations inside our containers. We even have the ability to define functions, triggers and stored procedures.
The support of writing queries directly inside the browser it's nice and I used with success to understand some concepts related to how to write queries. You can save the queries defined in the emulator, but we aware that they are saved inside Azure CosmosDB containers. This will generate an additional cost (~$0.77/day) but can be used as a common query repository by the team.
There is an option that allows us to reset all the data stored inside the emulator. Useful during development phases. There is also the ability to upload content from JSON, for cases when you need to prepare your local emulator with fresh data.
The speed and the performance of the emulator are not high, but it is more than enough for development and to play around. Locally I don't see an issue with lack of support of consistency level especially because you have only one instance - no replica.
To be able to activate emulators for Gremlin API, Cassandra API, Table API or MongoDB emulator you need to start the emulator from the command line and specify "/EmulatorMongoDBEndpoint" as the parameter for MongoDB. A similar thing you need to do if you want the emulator for the other APIs. Except for the SQL API that is fully supported, the rest of the APIs are partially supported.
The Docker support for the emulator is awesome and enable you to use the emulator in different and configurations.
Comments
Post a Comment