Jobs log datasource breaking change¶

Job type "populate" changing to "populateview"¶

Starting on August 1st, 2025, when you query the tinybird.jobs_log datasource and filter by job_type = 'populate', the query will stop returning results for new job entries. You will need to update your queries to use job_type = 'populateview' instead.

If your application currently monitors or processes jobs using the "populate" job type, you will need to update your code to handle "populateview" instead. This includes:

• Monitoring dashboards that track populate job performance
• Automated scripts that filter or process populate jobs
• Any alerting systems based on populate job status

We are making this change to align the Jobs API validation with the actual job system implementation and resolve inconsistencies between our documentation and API behavior. This change is part of our ongoing effort to standardize job type naming across all Tinybird systems.

Migration recommendations:

• Update your queries to use job_type = 'populateview' before August 1st
• Consider using job_type IN ('populate', 'populateview') during the transition period if you need to query both old and new job entries
• Test your monitoring and alerting systems with the new job type name

This change will go live on August 1st, 2025. We recommend updating your applications and monitoring systems well before that date to ensure a smooth transition. If you have any questions or concerns, reach out through our usual support channels.

Deployments now report the data that will be copied¶

The output of the tb deploy command now includes a section that reports the data to be copied during the deployment.

 Data that will be copied with this deployment:
---------------------------------------------------------------
| Datasource | Backfill type                                  |
---------------------------------------------------------------
| events     | From live deployment using Forward Query       |
| processed  | Using Materialized Pipe process_events         |
---------------------------------------------------------------

Or like this when no data will be copied:

* No data will be copied with this deployment

Skip backfilling a new Data Source¶

By default, when you create a Data Source that is the destination of a Materialized View, it is automatically backfilled with the results of the Materialized View's query. You can now skip this behavior by adding the following to the end of the new .datasource file:

BACKFILL skip

Observability of the memory consumption of your endpoints¶

Now you can track the memory consumption of your API endpoints with the new memory_usage column in tinybird.pipe_stats_rt. This new metric helps you identify inefficient queries and optimize endpoint performance.

The memory_usage field shows the memory consumption for each query in bytes. High memory usage can indicate:

• Complex operations that need optimization
• Large result sets that could be paginated
• Unoptimized queries that scan more data than necessary

Use this data to monitor and improve your endpoint efficiency. For example, you can find endpoints with the highest memory usage:

SELECT
    pipe_name,
    formatReadableSize(avg(memory_usage)) as avg_memory_usage,
    count() as request_count
FROM tinybird.pipe_stats_rt
WHERE start_datetime >= now() - INTERVAL 24 HOUR
GROUP BY pipe_name
ORDER BY avg(memory_usage) DESC

Learn more about monitoring endpoint performance in our performance analysis guide.

You can also visualize memory usage trends over time to identify patterns and spikes in memory consumption:

Improvements and bug fixes¶

• [All]: Improved consistency in job type naming across the Jobs API and internal systems.