> For the complete documentation index, see [llms.txt](https://docs.flopsar.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.flopsar.com/7/user-guide/observing-and-diagnosing/data-browser.md).

# Data Browser

The Data Browser is where you query the method invocations that agents have recorded and inspect them one by one. It is usually the first stop in a post-hoc investigation: you describe what you are looking for, Flopsar returns the matching calls, and from there you can open any call's full execution tree, read its parameters, or examine the exception it threw.

## Running a Search

Open the search form with the **Search** button in the top-right corner. Until you run a search the page shows an empty state — the browser does not load data on its own, because the recorded dataset can be very large and queries are always scoped.

<figure><img src="/files/f6pBMwMVUKHdQi99Gv9D" alt="" width="563"><figcaption></figcaption></figure>

The form has two layers. At the top you set the scope that always applies:

* **Date range** — the time window to search. This is the single most important constraint; keep it as narrow as the investigation allows.
* **Filter** — an optional predefined filter group. Selecting one applies a saved set of conditions and hides the manual fields below. Leave it empty to build a custom query (the form then shows a "Custom filter" divider).&#x20;

When no filter group is selected, you compose a custom query from these fields:

* **Root calls only** — restrict results to top-level (entry-point) invocations rather than every nested method. Useful when you want one row per logical request instead of the whole call graph.
* **Application** — the monitored application to search. This is required.
* **Instance** — a pattern matching the specific agent instance (host/process).
* **Thread** — the thread name; suggestions are offered as you type.
* **Class** — the fully qualified class name (suggestions appear after a few characters).
* **Method** — a specific method on the selected class.
* **Exception** — match only invocations that threw a given exception type.
* **Parameter conditions** — match on captured parameter values. Enter a **parameter key**, a **value** pattern, and add the condition with the **+** button. You can stack several conditions; each is listed below with its key, value, and match kind, and can be removed individually.&#x20;

Submit with **Search**. The matching invocations populate the results table.

## Reading the Results

Each row is one recorded invocation. The table shows:

* **Instance** — the agent instance, with the application name beneath it.
* **Thread** — the executing thread.
* **Time** — the start timestamp (time and date).
* **Method** — the method name; click it to open the execution tree (see below). The full signature is shown underneath.
* **Class** — the class, with its package beneath.
* **Parameters** — a preview of captured parameters; click to open the full parameter inspector.
* **Exception** — the exception type, if the call threw one (highlighted in red); click to view the exception details.
* **Duration** — total wall-clock time of the invocation.
* **CPU** — CPU time consumed by the invocation.&#x20;

Every column header is **sortable** (click to toggle ascending/descending), and columns can be **resized** by dragging their right edge; the reset button in the toolbar restores default widths.&#x20;

Two toolbar controls refine what you see without re-querying the server:

* **Filter invocations** — a quick text filter applied to the rows already loaded.
* **Preferred param key** — choose which captured parameter key is shown in the Parameters column, so the value most relevant to your investigation is visible at a glance.&#x20;

Results are paginated (20, 50, or 100 rows per page). When the server has more matching data than was returned in the first batch, the **Load more results** button in the header fetches the next batch.

## Drilling Into a Call

Three row-level actions take you deeper:

* **Method name → Call Stack Details.** Clicking a method opens the full method invocation tree for that call — the nested calls it made and the time spent in each. This is the primary bridge into the [*Method Execution Tree*](/7/user-guide/observing-and-diagnosing/method-execution-tree.md), documented on its own page.