Requests

Requests is the screen you open when cost looks odd, an error shows up, or someone reports a strange response. Every AI call that passed the gateway is kept here, so you narrow the list with filters and then open one call to find the cause. For the full process of narrowing a cost spike down to this screen, see Respond to a cost spike.

The list and the detail

Area	What you see
List	One request per row. Key fields include time, provider, model, user, tokens, cost, latency, and status. A request ABTO retried shows its last attempt, with a ↻ mark next to the status
Detail: conversation	The actual system and user messages and the response body. Failed calls show the raw error message
Detail: breakdown	Prompt versus completion tokens, input versus output cost
Detail: events	What the user who received this response actually did
Detail: attempts	Shown only for retried requests. Lists each attempt in order with its status code, error message, and provider response time; click a line to open that attempt

Failed rows get a red tint with the error message shown alongside. The token and cost breakdown makes anomalies like an abnormally large prompt stand out immediately. Latency is the time from ABTO receiving the request to sending the response, including any time spent on retries. The detail also shows the time the model provider took to respond (provider response time) on its own.

If the detail says “Logging is disabled or no messages were recorded”, no message bodies were kept for that call. Tokens, cost, and latency are still recorded all the same. Whether bodies are saved is set in Features and data in Settings.

Tools for narrowing the cause

Tool	Use
Filters	Conditions such as feature, model, variant, status (success or error), and range. The variant filter turns on once you pick a feature, because variants belong to a feature. If you arrived by clicking Overview’s error rate number, the list is already filtered to the requests the error rate counts (those that ended in a provider error, a timeout, a connection failure, or an ABTO internal error)
Sorting	The time and latency columns. Sort by latency to find the slowest calls first
Columns	Use the Columns button at the top right to choose which columns to show, and reorder them by dragging the ⋮⋮ handle or clicking it for a move menu. With a mouse, you can also drag a column name in the list. The time column stays first, and the column you are sorting by cannot be hidden. The setting is kept in this browser per user and workspace, and Reset to default restores the original layout
Automatic updates	Checks the latest first page every five seconds while the tab is visible. The first saved request and subsequent requests appear automatically; unchanged tables stay in place. Updates pause in hidden or offline tabs and resume immediately on return. Failed reads retain the list and retry with increasing delays. After you load older requests, the list and reading position stay in place and the button changes to Show latest when updates are available
Refresh / Show latest	Reloads the first page of data rows and updates the count and pagination while keeping the current range, filters, sorting, and table header. Current variant names update with the rows, and column widths and horizontal scroll position stay stable even when values become shorter. If the account or workspace changes in another tab, reopen the page before refreshing. A spinning button icon indicates the update; the existing list stays visible if it fails
Open in Testing	Loads this call’s request body (earlier conversation, tool calls and results, system prompt) into Testing exactly as recorded. Image and PDF attachments are shown as placeholders and are not included in a run