Traces, spans, run logs, cost tracking, and token budgets — understanding what your agent is doing in production.
Key vocabulary
Agent trace — the complete record of one agent run: all steps, tool calls, LLM calls, results.
Span — a single unit of work within a trace (one LLM call, one tool call).
Agent run log — structured log output from an agent run, for debugging.
LLM call cost — the token-based cost of each model API call within the run.
Token budget — a pre-set limit on total tokens the agent run may consume.
0 / 26 completed
1 / 26
An agent trace records:
Agent trace = the full observability record. Captures every LLM call, tool call, tool result, and reasoning step. Essential for debugging, cost analysis, and performance optimisation.
2 / 26
A span in agent observability is:
Span (from OpenTelemetry) = one atomic operation: one LLM API call, one tool execution. Spans have start time, end time, and metadata. Multiple spans compose a trace.
3 / 26
Why do production agent systems track LLM call cost per run?
LLM call cost tracking: agents make many calls per task and costs compound. Tracking enables: identifying expensive tasks, optimising prompts, setting cost budgets, and alerting on anomalies.
4 / 26
A token budget for an agent run is:
Token budget = total token allowance across all calls in one run. When reached, the agent must stop. Prevents runaway agents from consuming unbounded resources.
5 / 26
The agent run log is used to:
Agent run log = structured output designed for engineers to read. Good run logs include: step-by-step reasoning, tool calls with arguments, tool results, and cost-per-step data.
6 / 26
Reviewer: 'I'm seeing a lot of high latency in the agent's responses. The traces show multiple spans with durations exceeding 500ms during the intent recognition phase. Can you investigate and optimize this? Specifically, I'd like to see if we can reduce the number of calls to the external knowledge base.',
You (responding in a Slack message): 'Okay, I'm looking into it. The trace data suggests a potential bottleneck related to the query latency of the knowledge base service. It's worth investigating whether we can improve caching or explore alternative retrieval strategies.'
This question tests understanding of 'query latency' within the context of agent observability. While query latency *does* relate to external service calls (and is a key component of overall performance), the phrase specifically describes the time taken during intent recognition – the initial step in an agent's process. Options A and B are too broad, while option D misrepresents its use; it's crucial for identifying bottlenecks, not just general debugging.
7 / 26
PR Description: 'Agent run logs show frequent 'AGENT_REQUEST_TIMEOUT' errors during the user authentication flow. Initial investigation suggests the external identity provider is experiencing intermittent high latency. We need to monitor these timeouts closely and consider implementing circuit breakers to mitigate impact.'
Which of the following phrases best describes the *root cause* identified in this PR description, focusing on observability terminology?
The PR describes an 'AGENT_REQUEST_TIMEOUT' error linked to high latency from an external identity provider. This indicates a problem with *availability* and *external dependency performance*, which are key concerns for observability. The phrase 'transient network issue' most accurately reflects the situation – a temporary, but significant, degradation in service response time that's directly observable through agent run logs. Options A, C, and D misinterpret the symptom (timeouts) as something internal to the agent itself.
8 / 26
A developer observes a spike in 'AGENT_REQUEST_TIMEOUT' errors during peak usage. The agent's internal logs show unusually high numbers of requests to the 'Recommendation Engine' service. The traces reveal that many spans related to this service have extremely long durations (over 1 second).
Which of the following best represents the observability concept being highlighted in this scenario?
The situation describes a dependency between the agent and the 'Recommendation Engine'. 'Correlation ID tracking' is essential for tracing requests across multiple services to understand the end-to-end flow. The long spans directly demonstrate this dependency's impact on the overall agent response time – a key aspect of SLO monitoring, but correlation ID tracking provides the foundational observability concept in this example.
9 / 26
Reviewer: 'I'm seeing a lot of high latency in the agent's responses. The traces show multiple spans with durations exceeding 500ms during the intent recognition phase. Can you investigate and optimize this? Specifically, I'd like to see if we can reduce the number of calls to the external knowledge base.',
You (responding in a Slack message): 'Okay, I'm looking into it. The trace data suggests a potential bottleneck related to the query latency of the knowledge base service. It's worth investigating whether we can improve caching or explore alternative retrieval strategies.'
This question tests understanding of 'query latency' within the context of agent observability. While query latency *does* relate to external service calls (and is a key component of overall performance), the phrase specifically describes the time taken during intent recognition – the initial step in an agent's process. Options A and B are too broad, while option D misrepresents its use; it's crucial for identifying bottlenecks, not just general debugging.
10 / 26
PR Description: 'Agent run logs show frequent 'AGENT_REQUEST_TIMEOUT' errors during the user authentication flow. Initial investigation suggests the external identity provider is experiencing intermittent high latency. We need to monitor these timeouts closely and consider implementing circuit breakers to mitigate impact.'
Which of the following phrases best describes the *root cause* identified in this PR description, focusing on observability terminology?
The PR describes an 'AGENT_REQUEST_TIMEOUT' error linked to high latency from an external identity provider. This indicates a problem with *availability* and *external dependency performance*, which are key concerns for observability. The phrase 'transient network issue' most accurately reflects the situation – a temporary, but significant, degradation in service response time that's directly observable through agent run logs. Options A, C, and D misinterpret the symptom (timeouts) as something internal to the agent itself.
11 / 26
A developer observes a spike in 'AGENT_REQUEST_TIMEOUT' errors during peak usage. The agent's internal logs show unusually high numbers of requests to the 'Recommendation Engine' service. The traces reveal that many spans related to this service have extremely long durations (over 1 second).
Which of the following best represents the observability concept being highlighted in this scenario?
The situation describes a dependency between the agent and the 'Recommendation Engine'. 'Correlation ID tracking' is essential for tracing requests across multiple services to understand the end-to-end flow. The long spans directly demonstrate this dependency's impact on the overall agent response time – a key aspect of SLO monitoring, but correlation ID tracking provides the foundational observability concept in this example.
12 / 26
Reviewer: 'I'm seeing a lot of high latency in the agent's responses. The traces show multiple spans with durations exceeding 500ms during the intent recognition phase. Can you investigate and optimize this? Specifically, I'd like to see if we can reduce the number of calls to the external knowledge base.',
You (responding in a Slack message): 'Okay, I'm looking into it. The trace data suggests a potential bottleneck related to the query latency of the knowledge base service. It's worth investigating whether we can improve caching or explore alternative retrieval strategies.'
This question tests understanding of 'query latency' within the context of agent observability. While query latency *does* relate to external service calls (and is a key component of overall performance), the phrase specifically describes the time taken during intent recognition – the initial step in an agent's process. Options A and B are too broad, while option D misrepresents its use; it's crucial for identifying bottlenecks, not just general debugging.
13 / 26
PR Description: 'Agent run logs show frequent 'AGENT_REQUEST_TIMEOUT' errors during the user authentication flow. Initial investigation suggests the external identity provider is experiencing intermittent high latency. We need to monitor these timeouts closely and consider implementing circuit breakers to mitigate impact.'
Which of the following phrases best describes the *root cause* identified in this PR description, focusing on observability terminology?
The PR describes an 'AGENT_REQUEST_TIMEOUT' error linked to high latency from an external identity provider. This indicates a problem with *availability* and *external dependency performance*, which are key concerns for observability. The phrase 'transient network issue' most accurately reflects the situation – a temporary, but significant, degradation in service response time that's directly observable through agent run logs. Options A, C, and D misinterpret the symptom (timeouts) as something internal to the agent itself.
14 / 26
A developer observes a spike in 'AGENT_REQUEST_TIMEOUT' errors during peak usage. The agent's internal logs show unusually high numbers of requests to the 'Recommendation Engine' service. The traces reveal that many spans related to this service have extremely long durations (over 1 second).
Which of the following best represents the observability concept being highlighted in this scenario?
The situation describes a dependency between the agent and the 'Recommendation Engine'. 'Correlation ID tracking' is essential for tracing requests across multiple services to understand the end-to-end flow. The long spans directly demonstrate this dependency's impact on the overall agent response time – a key aspect of SLO monitoring, but correlation ID tracking provides the foundational observability concept in this example.
15 / 26
Reviewer: 'I'm seeing a lot of high latency in the agent's responses. The traces show multiple spans with durations exceeding 500ms during the intent recognition phase. Can you investigate and optimize this? Specifically, I'd like to see if we can reduce the number of calls to the external knowledge base.',
You (responding in a Slack message): 'Okay, I'm looking into it. The trace data suggests a potential bottleneck related to the query latency of the knowledge base service. It's worth investigating whether we can improve caching or explore alternative retrieval strategies.'
This question tests understanding of 'query latency' within the context of agent observability. While query latency *does* relate to external service calls (and is a key component of overall performance), the phrase specifically describes the time taken during intent recognition – the initial step in an agent's process. Options A and B are too broad, while option D misrepresents its use; it's crucial for identifying bottlenecks, not just general debugging.
16 / 26
PR Description: 'Agent run logs show frequent 'AGENT_REQUEST_TIMEOUT' errors during the user authentication flow. Initial investigation suggests the external identity provider is experiencing intermittent high latency. We need to monitor these timeouts closely and consider implementing circuit breakers to mitigate impact.'
Which of the following phrases best describes the *root cause* identified in this PR description, focusing on observability terminology?
The PR describes an 'AGENT_REQUEST_TIMEOUT' error linked to high latency from an external identity provider. This indicates a problem with *availability* and *external dependency performance*, which are key concerns for observability. The phrase 'transient network issue' most accurately reflects the situation – a temporary, but significant, degradation in service response time that's directly observable through agent run logs. Options A, C, and D misinterpret the symptom (timeouts) as something internal to the agent itself.
17 / 26
A developer observes a spike in 'AGENT_REQUEST_TIMEOUT' errors during peak usage. The agent's internal logs show unusually high numbers of requests to the 'Recommendation Engine' service. The traces reveal that many spans related to this service have extremely long durations (over 1 second).
Which of the following best represents the observability concept being highlighted in this scenario?
The situation describes a dependency between the agent and the 'Recommendation Engine'. 'Correlation ID tracking' is essential for tracing requests across multiple services to understand the end-to-end flow. The long spans directly demonstrate this dependency's impact on the overall agent response time – a key aspect of SLO monitoring, but correlation ID tracking provides the foundational observability concept in this example.
18 / 26
Slack Message from Sarah (a Senior Engineer): 'Hey team, we're seeing a surge in `AGENT_REQUEST_TIMEOUT` errors in the production logs. The agent is repeatedly failing to connect to the Recommendation Engine service during peak hours. Initial traces show significant delays when querying the engine's API. What's the *primary* thing you should investigate based on this message?
While all options *could* play a role, Sarah's message explicitly states 'timeout errors' and 'delays when querying the engine's API'. This strongly suggests a problem with the Recommendation Engine service itself – perhaps it's overloaded or experiencing issues. The other options are possible contributing factors but aren't the primary focus of the log data.
19 / 26
PR Description: 'Agent run logs show frequent 'AGENT_REQUEST_TIMEOUT' errors during the user authentication flow. Initial investigation suggests the external identity provider is experiencing intermittent high latency. We've added retry logic to the agent, but the errors persist. The agent traces reveal that many requests are being dropped before the retries even execute. What does this indicate about the current retry strategy?
If requests are being dropped *before* retries execute, it means the initial latency is so high that the retries aren't even having a chance to succeed. An overly aggressive retry strategy would only exacerbate this problem by continuously hammering the provider with failed requests. Exponential backoff is a good strategy but doesn't fix the underlying issue – the provider is still slow.
20 / 26
Standup Update from David (a Developer): 'I've been investigating the recent spike in `AGENT_REQUEST_TIMEOUT` errors. The agent logs show a significant increase in requests to the 'Sentiment Analysis' service during peak hours. Traces reveal that many of these requests are timing out while processing complex, multi-sentence user inputs. I'm considering increasing the timeout duration for this service but want to understand if there's an underlying issue with the sentiment analysis model itself.'
David's observation about 'complex, multi-sentence inputs' suggests the problem isn't simply a timeout issue; it's the *load* on the sentiment analysis model. Increasing the timeout duration would only mask the symptom and wouldn't address the root cause of the model struggling to process complex requests – this points to needing more resources or a more efficient model.
21 / 26
Code Review Comment: 'I'm seeing a lot of high latency in the agent's responses. The traces show multiple spans with durations exceeding 500ms during the intent recognition phase. Can you investigate and optimize this? Specifically, are there any assumptions being made about the user's input that could be causing unnecessary processing delays?
The question directly asks about potential assumptions in the user's input. This suggests there might be unnecessary pre-processing or feature extraction happening that could be contributing to the latency. Focusing on this aspect is a valuable debugging step before assuming problems with the core engine.
22 / 26
During a daily standup, Mark reports: 'We're observing increased `AGENT_REQUEST_TIMEOUT` errors when processing user queries related to product recommendations. The agent's metrics dashboard shows a consistent spike in these timeouts over the last hour. What is the *most* appropriate response to convey this issue to the broader team?',
This question tests understanding of communication within a development team. Option 2 is best because it acknowledges investigation without overreacting or assigning blame. Options A and B are too simplistic; option C implies an immediate escalation that may not be necessary at this stage. Option D requires further context.
23 / 26
Sarah (Senior Engineer) sends this Slack message: '@team, we're seeing a surge in `AGENT_REQUEST_TIMEOUT` errors. The agent is repeatedly failing to connect to the Recommendation Engine service during peak hours – specifically when users are browsing our new summer collection. Can anyone suggest potential causes?' Which of the following options represents the *best* immediate action Sarah should take?
This question focuses on triage and data gathering. Requesting traces is the most efficient way to diagnose the root cause of the issue. Options A dismisses the problem; B suggests premature action without investigation; and C inappropriately shifts responsibility.
24 / 26
You're reviewing a PR that introduces a new feature for the agent. The code includes logging statements designed to capture `AGENT_REQUEST_TIMEOUT` events. Which of the following best describes the *purpose* of these logs in this scenario?
The logs are being used for observability – to analyze the *patterns* of timeout events. This allows developers to understand when and why timeouts occur, which is crucial for identifying root causes and preventing future issues. Option A describes detailed timestamps; option C represents an unrealistic goal; and D misrepresents the log's function.
25 / 26
During a code review, Alex comments: 'I'm seeing elevated latency in agent responses, particularly during intent recognition. The traces show spans exceeding 600ms. Can we investigate the performance of the NLU model?' What does Alex *most* likely want to discuss?
Alex's comment directly references 'intent recognition,' a core function of an agent. Elevated latency in this phase strongly suggests a problem with the NLU model itself – its processing time is excessive. While other services could contribute to latency, Alex's focus is on the immediate performance bottleneck identified in the traces. The options regarding architecture and configuration are less relevant to the specific symptom described.
26 / 26
David (a Developer) reports: 'I've been investigating the recent spike in `AGENT_REQUEST_TIMEOUT` errors. The agent logs show a significant increase in requests to the 'Sentiment Analysis' service during peak hours when users are asking about current events – specifically, news headlines related to financial markets.' What is David *primarily* communicating about?
David is describing a correlation between peak hours and an increase in sentiment analysis requests. This suggests user behavior (asking about news) is driving the increased load on that service. The other options are less likely based on the information provided.
What will I practice in "Agent Observability Vocabulary | Coders Lingo"?
This is an AI Agents Language exercise set. It walks through 26 scenario-based multiple-choice questions built around real usage of AI Agents Language terminology that IT professionals encounter on the job.
Is this exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is free to complete with no account, sign-up, or paywall.
How many questions are in this exercise?
This set contains 26 questions. Each one shows immediate feedback and a detailed explanation after you answer, so you learn the correct usage right away rather than waiting for a final score.
Do I need prior experience to complete this exercise?
No prior experience is required. Each question includes a full explanation covering the reasoning behind the correct answer, so the exercise itself teaches the AI Agents Language vocabulary as you go.
Can I retry the exercise if I get questions wrong?
Yes — use the "Try again" button on the results screen to reset your answers and go through all the questions again. There is no limit on attempts.
Is my progress saved?
Your answers and score for the current session are tracked in the browser as you go. No account or login is needed, and there is nothing to install.
What if I don't understand a term used in a question?
Read the explanation shown after you answer each question — it breaks down the correct term in plain English with a real-world example. You can also check the site Glossary for quick definitions.
How is this different from reading a blog article on the topic?
Exercises like this one are interactive drills that test and reinforce specific vocabulary through multiple-choice questions, while blog articles explain concepts in prose. Practising here after reading builds active recall, not just passive recognition.
Where can I find more AI Agents Language exercises?
See the AI Agents Language exercises hub for the full set of related pages, or browse all exercise categories from the main Exercises index.
Can I use this exercise to prepare for a technical interview?
Yes — AI Agents Language vocabulary comes up often in technical discussions and interviews. Pair this exercise with our dedicated Interview Preparation section for role-specific practice.