Skip to content

Start the agent when a collector is used - #280

Open
unflxw wants to merge 1 commit into
mainfrom
start-agent-in-collector-mode
Open

Start the agent when a collector is used#280
unflxw wants to merge 1 commit into
mainfrom
start-agent-in-collector-mode

Conversation

@unflxw

@unflxw unflxw commented Aug 27, 2026

Copy link
Copy Markdown
Contributor

Setting collector_endpoint replaces the agent, so host metrics, NGINX
metrics, StatsD metrics and environment metadata are not reported at
all. The collector implements none of them for the application's host:
it can report host metrics, but for the host it runs on, which is
another machine.

The agent now runs alongside the collector, and only the trace, metric
and log data moves to the collector. Its OpenTelemetry listener is
turned off there, because that data goes to the collector instead and
the two default to the same port. The agent no longer decides whether
AppSignal starts, so collector mode keeps working where the agent is
unavailable, and stopping it no longer waits for it to flush, because
what it holds is a minute of host metrics.

Most agent options apply again, leaving only those that configure how
it handles trace data to warn that they are ignored.

AppSignal for Ruby turns the agent's OpenTelemetry listener off in appsignal/appsignal-ruby#1595.

The documentation for this change is in appsignal/appsignal-docs#199.

@backlog-helper

backlog-helper Bot commented Aug 27, 2026

Copy link
Copy Markdown

Hi @unflxw,

We've found some issues with your Pull Request.

  • This Pull Request is missing labels. Please add labels to help identify types of Pull Requests. - (More info)
  • This Pull Request is missing reviewers. Either convert this Pull Request into a draft or ignore this rule by adding [skip review] to your Pull Request body. - (More info)

New issue guide | Backlog management | Rules | Feedback

@unflxw
unflxw marked this pull request as ready for review August 27, 2026 16:30
@backlog-helper

Copy link
Copy Markdown

Hi @unflxw,

We've found new issues for this Pull Request. Please see the main comment on this issue for a list of all current warnings. This comment will not be updated to reflect resolved warnings.

  • This Pull Request is missing reviewers. Either convert this Pull Request into a draft or ignore this rule by adding [skip review] to your Pull Request body. - (More info)

New issue guide | Backlog management | Rules | Feedback

Setting `collector_endpoint` replaces the agent, so host metrics, NGINX
metrics, StatsD metrics and environment metadata are not reported at
all. The collector implements none of them for the application's host:
it can report host metrics, but for the host it runs on, which is
another machine.

The agent now runs alongside the collector, and only the trace, metric
and log data moves to the collector. Its OpenTelemetry listener is
turned off there, because that data goes to the collector instead and
the two default to the same port. The agent no longer decides whether
AppSignal starts, so collector mode keeps working where the agent is
unavailable, and stopping it no longer waits for it to flush, because
what it holds is a minute of host metrics.

Most agent options apply again, leaving only those that configure how
it handles trace data to warn that they are ignored.
@unflxw
unflxw force-pushed the start-agent-in-collector-mode branch from 213843e to c9b9c61 Compare August 28, 2026 12:08
@unflxw
unflxw changed the base branch from stop-probes-when-stopping to main August 28, 2026 12:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant