The Spark History Server is the standard upstream UI for Spark applications that have already completed. Where the Spark UI in Spark UI Access how's the application currently running on your kernel, the History Server is where you go after the kernel has gone — to inspect a past run, compare two runs, or verify a process result.
Getting there
Point your browser at https://{your-syntasa-host}/spark-server. The same URL is also accessible from the portal navigation.
What you see
The History Server's applications list shows every Spark application you have run that has reached a completed state — successful runs, failed runs, aborted runs. Each entry has a name, an application ID, the start and end times, and the duration. Clicking into an application opens the same Jobs / Stages / Storage / Environment / Executors / SQL views you get in the live Spark UI, except the data is read from the captured event log rather than a running driver.
Anything that ran a Spark session shows up here: interactive notebook runs, Notebook Process executions, and Spark code processes. The list is the union of all of them, sorted by completion time.
You only see your own jobs
The History Server filters the application list by the user who ran each application. You see the applications you ran; you do not see applications other users ran, even if you are members of the same workspace. This keeps the list focused on your own work.
Users with the System Admin role see every application across the platform — they need that view to operate the system and help users debug. If you are a System Admin and you see other users' jobs, that is by design.
Common uses
- Inspect a past notebook run that errored — drill into Stages and SQL to find the failing operation.
- Compare two runs of the same notebook (yesterday vs today) to spot a performance regression.
- Verify a Notebook Process actually executed and check the Spark configuration it ran with.
- Look at the Spark UI of an application that finished long ago, without re-running anything.
Related sections
- Spark UI Access. For the application currently running on your kernel. The History Server is the equivalent UI for applications that have already finished.
- Notebook Process (Jobs). Process runs land in the History Server when they complete. Use the History Server to compare a successful scheduled run to a failed one, or to inspect a run from before you opened the notebook today.