Closed szaffarano closed 1 week ago
Pinging @elastic/security-solution (Team: SecuritySolution)
To update your PR or re-run it, just comment with:
@elasticmachine merge upstream
Status | Branch | Result |
---|---|---|
✅ | 8.14 |
Note: Successful backport PRs will be merged automatically after passing CI.
Please refer to the Backport tool documentation
Summary
TelemetryLogger
now includes always by default some basic data (like cluster uuid), so it is highly recommended to start using it when possible (instantiate it using thenewTelemetryLogger
function)endpoint-meta-telemetry
endpoint-meta-telemetry
to simplify its logic and also improve error handling. Based on the tests, I had to add or change how we managed errors for some (edge) cases. Also, the errors now log more information introducing fields for relevant data instead of using a nonstatic string (e.g., instead ofRunning task: ${taskId} [last: ${taskExecutionPeriod.last} - current: ${taskExecutionPeriod.current}]
, nowtaskId
and theexecutionPeriod
are searchable fields and not part of the log message, which becomes a "static string"). Regarding error handling, the goal is to send as much information as possible and not abort the execution when we encounter weird (but still feasible) errors.endpoint-meta-telemetry
using the already defined feature flag, just in case we need to disable it through a configuration artifact. This change deprecates the current way we split the events to send them into batches (although if we disable the feature flag, it still uses this approach).