- OneClick Enhancements - Adding your inputs with just one click is getting even easier!
- We've added OneClick support to our Microsoft SQL Server and Oracle inputs as well!
- Tables that are created from data with a strict schema like a transactional database now get their primary and dist keys set automatically.
- The primary key (or keys) for the Redshift table will be retrieved from the original table's schema.
- The distribution key will be set to the first primary key of the original table.
- New Input: Elasticsearch - Your super flexible document store and search engine can now make it to your data warehouse! We've added our new Elasticsearch input, which allows you to define an index (and customized retrieval query, if you wish), and have that index replicated at the interval of your choice to your data warehouse. Check out our Elasticsearch doc for more info!
- Load to Multiple Schemas - Have you ever wanted to split your data into different schemas in your target database? Well, if you're a user of our
alooma.pyPython interface, now you can! Just define the schema of your choice when you set your mapping, and we'll write that event type directly into that schema. Note that if you don't define a schema, we'll write to the default one you configured when you connected your output.
- Splitting Code Engine Code into Modules - More productivity for our Code Enginer programmers and
alooma.pyusers - the ability to split up code into modules! Gone are the days when all your code had to be in one never ending file.
set_transform(code_str, module_name='name')allows you to set as many named modules as you like. If you don't specify a module name, the code will be loaded as the
mainmodule which contains your entrypoint: the
def transform(event):function. If you don't specify code, it will delete the specified module.
get_all_transformsreturns the list of modules you have defined in your system.
get_transform(module_name='main')will return the code defined for any module (defaults to your main code).
- If you have any questions about modules,
alooma.py, or anything else, don't hesitate to reach out to us at [email protected].
- Flexible Database Replication Key - Where once our incremental dump/load database replication needed all tables to have an explicit primary key as part of the table, we've now relaxed that requirement. Any column which acts like a primary key, even if it's not defined as such, may be configured in Alooma to serve as the records' unique identifier! Since this is an advanced configuration, please contact our support to take advantage of it.
- An issue with our reported metrics was causing incorrect charts and data to be shown in our dashboard and email digests. We found the culprit, and all should be back to normal.
- We identified a few key places where we weren't playing nice with Firefox. Now Firefox is back to being completely supported!
- Fixed a minor bug which interfered with connections to PostgreSQL servers on Heroku
- Fixed a bug which caused the S3 Input to misbehave when encountering binary files
- Fixed an issue which skewed incremental replication of tables with
nullvalues in the
updated_atcolumn. This bug was encountered with tables which didn't have an
updated_atcolumn, and were altered to include it. The resulting values would be
nullfor all existing records, and timestamps only for new records.