Skip to content

[BUG] new source:intake tag for ntp.offset is a breaking change #48789

@thom-vend

Description

@thom-vend

Agent version

7.77.2

Bug Report

The new feature [NTP] generate ntp.offset metric based on intake's HTTP response , PR: #45664 from @ian28223 is actually a breaking change, while it's nice to have a way to detect drift behind firewalls:

  • It created a lot of false positives on our monitors.
  • Sharing the same metric name ntp.offset is misleading; a name like ntp.http_fallback_offset would have been clearer.
  • Also, it doesn't account for network latency, unlike NTP, showing much higher values.

We had some unusual NTP drift reported above, near -900 ms since March 19, when we started having hosts with this new feature, probably because my team is using the us5 Datadog endpoint while running systems in AWS us-west-2.

Since reverting this would also be a breaking change, maybe a documentation update would be the bare minimum? (https://docs.datadoghq.com/integrations/ntp/?tab=host) + nothing mention this in the config file

This new feature has been published as an enhancement under 7.77.0

Anyway, thanks for the new feature and have a nice day :)

Reproduction Steps

Upgrade your agent to 7.77 and above while having existing monitors on ntp.offset

Agent configuration

default ntp configuration

Operating System

6.8.0-1050-aws #53-Ubuntu SMP Fri Mar 6 11:05:48 UTC 2026 aarch64 aarch64 aarch64 GNU/Linux

Other environment details

AWS, us-west-2, EC2 hosts

Metadata

Metadata

Assignees

Labels

oss/0External contributions priority 0pendingLabel for issues waiting a Datadog member's response.team/agent-runtimes

Type

No type
No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions