- OneClick Inputs - The future is here - OneClick setup of inputs!
- Gone are the days of always needing to map your newly created inputs, even if the mapping is straightforward. We've given you the option of creating your input with just one click (aka OneClick), and we'll map all event types generated by that input, create the associated tables, and your data will flow directly to your target database with no additional work required.
- It wouldn't be truly one click if future event types weren't auto-mapped (and tables created) as well, so they will be too!
- Of course, any new fields on existing event types for this input will be automatically mapped and created as well.
- To use OneClick create an input and keep OneClick as the selected "Mapping Mode" option.
- To manually map each event type, as today, just choose the Customized option.
- Any of your existing inputs will be set to Customized, and will continue working as is. In a future release you'll be able to turn existing inputs to OneClick mode as well.
- For users who utilize the Mixpanel or Segment inputs, notice that these inputs when created will be automatically split into event types, according to the original Mixpanel or Segment event type.
- New Replication Types: Full and Incremental Dump/Load - Because "one size fits all" just doesn't ring true when it comes to databases, we've added 2 new replication types to give you a broader range of options for how you'd like to replicate your tables.
- With full dump/load replication a full dump of the database is performed periodically, which replaces the existing database. It's simple and works best for small tables, but can only be run every few hours as it's resource heavy on your database.
- With incremental dump/load replication the database is queried periodically for updates since the last query, and updates are replicated. This works well for larger tables with much lower resource usage, but requires database having an incrementing numeric column for tracking the latest updates (e.g. updated_at), and doesn't capture deletes.
- We added these replication options to our MySQL and PostgreSQL inputs, so you can pick the replication type that best suits your needs!
- For any questions on replication options please reach out! We're here as always at firstname.lastname@example.org.
- New Inputs: Microsoft SQL Server, Oracle - In our valiant quest to make sure all of your data makes it to your target database, we've added some high-value new input types. You can connect to your Microsoft SQL server or Oracle database with either of our new replication types: full dump/load replication or incremental dump/load replication. Party on with all your data!
- Better Security for SDKs, Webhooks, and Others - As you've probably seen with our recent updates, we don't take security lightly. That's why were making some changes to the way you send events to Alooma.
- Now all inputs that push events to Alooma push to a single new, load balanced and secured Alooma endpoint: https://inputs.alooma.com.
- You must include a token with every event sent to the new endpoint - all of our SDKs support it.
- To get a token, create an input, give it a name, and a token will be generated for you.
- Note that you will get a different token for each labeled input type.
- Click Finish (yes, OneClick works here too!) and you'll see the new node created for you in the plumbing screen, waiting for events with this token to arrive at the endpoint.
- Expanded Code Engine Library - You asked for it, you got it! Support for new and expanded libraries is now in place:
- We've improved our auto-mapping of types even more to now also map to bigint and also have better timestamp and string detection.
- It's easier than ever to navigate back and forth while creating an input as we've added a location breadcrumb to the add input flow.
- We've expanded our S3 input support to also support unzipping
- Our previous 1MB event limit has been upped to 3MB by popular demand (though we still recommend sending events no larger than a few KB as large events may increase pipeline lag time).
- We've improved our Google Analytics input by giving you a notification if an input you defined returns an empty report, as well as pulling at least one day's worth of data by default.
- In our quest to simplify simplify simplify, we've gone ahead and removed the server ID field from MySQL inputs.
- We've improved the Run time feedback on running tests on sample data in the Code Engine to show a more accurate event transform processing time.
- Our auto-mapper was not mapping fields in metadata if all other fields were already mapped, so we remedied that.
- We removed a confusing "Code changed successfully" message that was being emitted on upgrades, despite the code not having changed.
- We fixed a meaningless error message that MongoDB was emitting when it hit authentication errors.