> ## Documentation Index > Fetch the complete documentation index at: https://hellotars.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Monitor tool calls > Read every call a toolkit or CodeKit made on its Logs view, filter the calls, and open one call to see its input, output, and console. Use the **Logs** view to find a failing call and see what was sent and returned. For trends across all your tools, see the **Tools and integrations** card in Analytics. See [Read the Analytics cards](https://hellotars.com/docs/guides/analytics/overview-tab.md#tools-and-integrations). The Logs list shows execution attempts. A failed attempt followed by a successful retry can appear as two rows. The summary counts them as one call, using its final outcome, so a failed row does not always add to the summary’s **Failed** count. ## Before you start Before you start, sign in to [the dashboard](https://dashboard.hellotars.com) and select the organization you want to work in. * You need a connected toolkit or a CodeKit toolkit. See [Connect a toolkit](https://hellotars.com/docs/guides/tools/connect-a-toolkit.md) and [CodeKit overview](https://hellotars.com/docs/guides/tools/codekit-overview.md). * The **Logs** view of any toolkit is shown only to Admins, Members, and Auditors. ## Open the Logs view **1. Open Tools, then use the toolkit’s More actions menu and select Open** For a connected toolkit, pick the account in the account picker. Logs show the calls from that account. **2. Switch the view from Configure to Logs** The view switcher sits beside the toolkit name. **3. Pick a time range** Choose **Last 24 hours**, **Last 7 days**, **Last 30 days**, or **Last 90 days** from the time-range dropdown. For specific dates, open the adjacent **Custom** date picker and select the start and end dates. The dropdown then reads **Custom range**. [https://dashboard.hellotars.com](https://dashboard.hellotars.com) ![Order Lookup Logs for the last 7 days with the Calls, Success rate, Failed, and Typical duration summary, the Calls per day chart, and the call list with Live tags](https://hellotars.com/docs/docs/guides/tools/images/tool-call-logs.png) ## Retained history Calls are available for **30 days on Free**, **90 days on Premium**, and **365 days on Enterprise**. Date presets outside your plan’s window are disabled, and a custom range cannot recover older calls. CodeKit console output is kept for **30 days on every plan**. A call can still be available after its console output has expired. The information icon beside the date controls shows both retention periods. See [Data retention](https://hellotars.com/docs/guides/security/data-retention.md). ## Read the summary The summary covers at most the **most recent 5,000 calls** within the selected, retained range. When results are partial, the help indicator on **Calls** explains that the numbers cover only the most recent 5,000 calls. Narrow the range before treating the counts or success rate as complete. This limit applies to the summary, not the number of attempt rows you can page through. | Element | What it shows | | - | - | | **Calls** | Calls included in the summary, counting retries once. | | **Success rate** | The share of calls that succeeded. | | **Failed** | Calls that failed. | | **Typical duration** | The median call duration. Half of the calls were faster. | | **Calls per day** | Succeeded and failed calls per day. | | **By action** | Calls per action, busiest first. | Some parts of the summary appear only in certain cases: * **Calls per day** appears when the range covers more than one calendar day. * **By action** needs more than one action and no action filter. * With **Mode** set to **CodeKit**, the summary is replaced by the note **Summary covers agent, flow step and test calls.** ## Filter the calls **Search action, agent or step** filters the list as you type. The filters narrow it by **Status**, **Mode**, and **Action**. **Component** appears when a component made calls in the range, such as the Meeting Scheduler. **Mode** separates calls by where they came from: **Agent**, **Workflow**, **Component**, **Test run**, and **CodeKit**. Each row carries a **Live** or **Test** tag. Hover it to see where the call came from: | Tag | Source | | - | - | | **Live** | A live call from an agent, from a flow step, or from a component. | | **Test** | A test chat in the dashboard, a run from the CodeKit **Test** tab, or a step test in the builder. | A test chat call links to the conversation in the test view. The list loads in pages. If a search finds nothing, select **Load more** to search the remaining calls in the range. [https://dashboard.hellotars.com](https://dashboard.hellotars.com) ![Order Lookup call history filtered to failed calls, with the search field, the Status: Failed chip, date controls, table headers, and durations](https://hellotars.com/docs/docs/guides/tools/images/tool-call-list.png) ## Open one call Select a row to open its details. The examples use **Get order status** from [Write a custom action](https://hellotars.com/docs/guides/tools/write-a-custom-action.md): `ORD-1042` returns a shipped order, while `ORD-9999` returns a **No order found** error. | Field | What it shows | | - | - | | **Agent**, **Step**, **Conversation** | Links to the agent, to the step on the builder, and to the conversation. | | **Source** | Shown only when a component made the call. It names the component, followed by **(component)**. | | **Role** | The part the action plays for a component, such as booking for the Meeting Scheduler. | | **Attempts** | Which attempt this row is, out of the maximum allowed, such as **1 of 3**. | | **Call ID** | The identifier of the call. | | **Input** and **Output** | The payloads, in read-only editors. | | Console | Console output for a CodeKit action, with times relative to the start of the run. Hover a time to see the clock time. | A payload or the console can be hidden in two cases: * The toolkit's **Log privacy** setting redacts payloads. They show as `[redacted]`, and the console reads **Console output is hidden by this tool's log privacy setting.** See [Redaction of tool execution logs](https://hellotars.com/docs/guides/security/data-protection.md#redaction-of-tool-execution-logs). * The call went through a CodeKit scheduling integration, for availability or for a booking. Tars does not keep their input and output, so both panels show `[redacted]`. ## Inspect a failed call Open a failed row to read its error alongside the call details. Check the action and inputs before retrying. ## Verify Use a read-only action in a test chat, then open **Logs**. The call appears at the top with a **Test** tag. ## Related pages * [Enable and use actions](https://hellotars.com/docs/guides/tools/enable-and-use-actions.md) * [Write a custom action](https://hellotars.com/docs/guides/tools/write-a-custom-action.md) * [Read the Analytics cards](https://hellotars.com/docs/guides/analytics/overview-tab.md) * [Book meetings with the Meeting Scheduler](https://hellotars.com/docs/guides/building/gambits/meeting-scheduler.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.