--- title: "Monitor agents: Activity, Deliveries, errors, success rate and credits" description: "Read the Agents list columns and the Activity and Deliveries tabs to see what each agent did, what failed and what it cost." url: https://help.k3x.ai/en/articles/80-monitor-agents-activity-deliveries-errors-success-rate-and-credits collection: "AI Agents" updated: 2026-10-04 language: en --- # Monitor agents: Activity, Deliveries, errors, success rate and credits Read the Agents list columns and the Activity and Deliveries tabs to see what each agent did, what failed and what it cost. The **Agents** tab is your control room. The list shows status, triggers, errors, success rate and credit spend per agent. Click an agent for the **Activity** tab (every run) and the **Deliveries** tab (every outbound webhook). ## The Agents list Filter tabs at the top: **All**, **Running**, **Paused**, **Drafts**. Columns: - **Name**: click to open the agent. - **Status**: Running (active and will fire), Draft (never activated), Paused (stopped by a person). - **Triggers**: the events that start it, such as Deal Updated, New Company, Webhook. Empty means on-demand only. - **Webhook**: toggle showing whether the agent's inbound webhook URL is switched on. - **Errors**: count of recent runs that ended in an error. Anything above zero is worth a click. - **Success Rate**: share of recent runs that completed without an error. It measures whether the agent finished its steps, not whether the company replied or the deal closed. - **Credits (7d)**: credits this agent spent in the last seven days. Use it to find the agent that is eating your pool. - **Creator**: who built it. - **Last Execution**: when it last ran. A Running agent with an old Last Execution has a trigger that is not firing. ## Activity tab Open an agent and click **Activity**. Each run shows the record it worked on, when it started, which step it reached, what the AI Analyzer observed (reply, click, booked meeting, no response) and the outcome. Open a run to read the full log: what the agent saw, what it decided and what it did. This is where you learn why an agent "did the wrong thing": almost always the goal allowed it. Agent actions also appear on each record's timeline. Filter the timeline by **System** for status changes and field updates, or by **Emails** and **SMS** for messages the agent sent. ## Deliveries tab If the agent has a step that sends a webhook to another tool, every attempt is listed on **Deliveries** with its destination, time and status. A failed delivery means the receiving URL did not accept the request; check that it is public HTTPS and that the receiver is on. ## Pause Agent The button at the top of the agent page toggles between **Pause Agent** and **Activate**. Pause when you want to edit a goal safely, when an agent is spending more than expected, or when you are about to import a large list and do not want New Company to fire a thousand times. Runs that are mid-step finish that step and then wait. Activate to resume. ## A weekly five-minute check 1. Sort by **Errors**. Open anything above zero and read the last failed run. 2. Sort by **Credits (7d)**. Make sure the top spender is the one you expect. 3. Scan **Last Execution** on Running agents. Stale dates mean a trigger or webhook sender stopped. 4. Glance at **Success Rate**. A drop usually follows a renamed stage or deleted custom field. 5. Check **Usage this cycle** in your profile and confirm **Auto-recharge** is on in **Settings → Billing** if you rely on agents running overnight. ## Notifications Email notifications for Critical Alerts and the Daily Digest, and in-app notifications for Company Updates and Pipeline Changes, are set in **Settings → Notifications**. Some of the emails you receive about an agent are notifications, not messages the agent sent to customers. ## Related - Fix an agent that does not run - Notifications: what K3X sends and how to control it - Your 7-day free trial and the Adaptive plan - Webhook-triggered agents