feat: set --max-poll-interval-ms flags to Sentry/Snuba consumers

[aldy505]

Separated from #4313

Also introducint a healthcheck for snuba-replacer

[sentry]

Bug: The snuba-replacer service is missing the --health-check-file flag in its command, which will cause its healthcheck to fail permanently.
_{Severity: MEDIUM}

Suggested Fix

Add the --health-check-file /tmp/health.txt flag to the command for the snuba-replacer service in docker-compose.yml to ensure the healthcheck file is created.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: docker-compose.yml#L373-L379

Potential issue: The `snuba-replacer` service is configured with a healthcheck that
depends on the existence of the file `/tmp/health.txt`. However, the service's `command`
is missing the `--health-check-file /tmp/health.txt` flag. This flag is necessary for
the process to create the health check file. Without it, the healthcheck will
consistently fail, and after the `HEALTHCHECK_FILE_START_PERIOD` of 600 seconds, Docker
will mark the container as permanently unhealthy.

_{Did we get this right? 👍 / 👎 to inform future reviews.}

[sentry]

Bug: The documented default for SENTRY_KAFKA_MAX_POLL_INTERVAL_MS in .env is 30000, but the docker-compose.yml default is 300000.
_{Severity: LOW}

Suggested Fix

Align the documented default value in the .env file with the actual default value used in docker-compose.yml. Change the comment in .env to reflect the 300000 ms default.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: .env#L51-L52

Potential issue: There is a mismatch between the documented default value for
`SENTRY_KAFKA_MAX_POLL_INTERVAL_MS` and its implementation. The comment in the `.env`
file states the default is `30000` ms, but the `docker-compose.yml` services use a
default of `300000` ms. This 10x difference could mislead users who rely on the
documentation, potentially causing them to set a much shorter interval than intended,
which could lead to the Kafka rebalancing issues this setting is designed to mitigate.

_{Did we get this right? 👍 / 👎 to inform future reviews.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 55c8540. Configure here.}

[cursor]

Replacer healthcheck missing heartbeat file

Medium Severity

snuba-replacer now uses the shared file-based healthcheck that succeeds only when /tmp/health.txt exists, but its replacer command never passes --health-check-file /tmp/health.txt like the other Snuba Kafka consumers. After the start period, Docker will mark the service unhealthy even while the replacer keeps running.

 

^{Reviewed by Cursor Bugbot for commit 55c8540. Configure here.}