I have both datadog_token and datadog_region set b...
# ask-ai
f
I have both datadog_token and datadog_region set but i see no logs being forwarded into the datadog instance - is there any open issue here? anything else i have to configure?
✅ 1
t
Thanks for reaching out, Rohit. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
Thanks — I’m sorry you’re not seeing logs in Datadog. I can help troubleshoot. A few quick checks so I can narrow this down: • Are your agents deployed to LiveKit Cloud or are they self‑hosted? (Automatic Datadog forwarding only works for LiveKit Cloud agents.) • Can you confirm the secrets were injected into the agent deployment (not just stored elsewhere) as
datadog_token
and
datadog_region
? • Do you see logs locally from the agent (try
lk agent logs --id <agent-id>
)? If yes, do those logs include any Datadog/forwarding errors? • Can you share the Datadog region value you set (e.g.,
us
or
eu
) and whether the token has the expected ingest/logs permissions? What I can tell you now: • There are no known service-wide outages for Datadog log forwarding. • If your agents are self‑hosted, automatic forwarding is not available and you’ll need to configure log forwarding yourself. • If the agent is in LiveKit Cloud and secrets are correctly injected, the next checks are agent log output and token/region values. Reply with the answers above (agent type, confirmation of secret injection,
lk agent logs
output or errors, and the Datadog region/token scope) and I’ll give the next specific steps. Sources: Where to find your Agent Logs | LiveKit | Where to find your Agent Logs | LiveKit | Agent deployment overview | LiveKit Documentation | Data Processing Addendum | LiveKit, with interactive elements
f
1. agents are deployed to livekit cloud. 2. yep the secrets can be seen from teh livekit UI. 3. yes i can see the logs. 4. eu.
t
Thanks — helpful details. Next steps to narrow this down: 1. Check agent logs for forwarding errors — search for “datadog”, “dd-logs”, or any HTTP/connection errors in
lk agent logs --id <agent-id>
. Share any matching error lines you find. 2. Verify the Datadog token has Logs Intake permissions (ingest API key for EU site). Confirm it's not an API key for metrics-only. 3. Confirm region mapping: you set
datadog_region=eu
(correct for Datadog EU). Double-check your Datadog intake endpoint accepts EU keys (e.g.,
http-intake.logs.datadoghq.eu
). 4. Restart the agent deployment after you updated secrets to ensure the runtime picked them up. 5. If you see logs locally but nothing in Datadog, try a simple test log line (e.g.,
log().info("datadog-test")
) and watch for any forwarding errors immediately after. Please paste any Datadog-related error lines from the agent logs and confirmation about the token type/permissions and I’ll give the next specific steps. Sources: Where to find your Agent Logs | LiveKit | Agent deployment overview | LiveKit Documentation | Where to find your Agent Logs | LiveKit, with interactive elements
f
i did all this - there is literally no error here. also what's surprising is that as i do a call, lk agent logs itself doesn't show me any logs.
t
Thanks — I hear your frustration. You mentioned no errors and that
lk agent logs
shows nothing during a call. A few clarifications and targeted checks so I can help further: • When you say “lk agent logs itself doesn't show me any logs,” do you see any logs from that agent at all after the call ends? (Agent observability uploads session logs only after a session completes — live session logs may not appear during the call.) • Please share the agent ID you’re testing with so I can look at its runtime status. While you confirm those, please do the following if you haven’t already: 1. After finishing a call, run:
lk agent logs --id <agent-id>
and paste any Datadog-related lines or any lines at all (even timestamps). 2. Add an explicit test log in the agent code before/after the call (e.g.,
log().info("datadog-test-start")
and
log().info("datadog-test-end")
) and check logs after the session ends. 3. Verify agent log level includes INFO/DEBUG so those test lines are emitted. 4. Confirm you restarted the agent after secrets were set (secrets are injected at runtime). 5. Confirm Datadog ingestion: does your Datadog account show any logs from other sources for the
eu
intake? This rules out an account-side ingestion issue. Send the agent ID plus the post-session
lk agent logs
output (or any Datadog-related log lines). Once I have that I’ll give the next specific steps. , with interactive elements