Skip to content

Latest commit

 

History

History
96 lines (67 loc) · 5.09 KB

KnownIssues.md

File metadata and controls

96 lines (67 loc) · 5.09 KB

Known Issues

Changes to collector properties are not applied

After running the collector for the first time, changes to collector properties (e.g. collector description, category, fields) are not applied.

To work around this, you need to delete the existing collector registration and register the collector again. To do this, you need to do two things:

  • Remove the collector in Sumo Logic UI

    • Log in to your Sumo Logic UI
    • Go to Manage Data - Collection
    • Find your collector
    • Click Delete on the right-hand side of the collector
  • Delete local collector registration file in ~/.sumologic-otel-collector/.

After that, the collector will register on next run.

If you delete the collector in the UI but not delete the local registration file, the collector will fail to start - see Collector fails to start when deleted from UI.

On the other hand, if you only delete the local registration file and do not delete the collector in the UI, a new collector will be created with current timestamp as a suffix, to prevent overwriting the existing collector.

Collector fails to start when deleted from UI

After successful registration of collector, if you delete the collector in Sumo Logic UI, the collector will fail to start on next run. The error message is similar to the below:

2021-08-24T10:52:38.639Z  error  [email protected]/extension.go:373  Heartbeat error  {"kind": "extension", "name": "sumologic", "collector_name": "<your-collector-name>", "collector_id": "0000000001A2B3C4", "error": "collector heartbeat request failed, status code: 401, body: {\n\"servlet\":\"rest\",\n\"message\":\"Could not authenticate.\",\n\"url\":\"/api/v1/collector/heartbeat\",\n\"status\":\"401\"\n}"}
github.com/open-telemetry/opentelemetry-collector-contrib/extension/sumologicextension.(*SumologicExtension).heartbeatLoop
  github.com/open-telemetry/opentelemetry-collector-contrib/extension/[email protected]/extension.go:373

To work around this, delete the local collector registration file at ~/.sumologic-otel-collector/. The collector will re-register on next run.

Enabling clobber property re-registers collector on every restart

If you set the extensions.sumologic.clobber property to true, a new collector registration that replaces the previously existing registration will be created on every run of the collector.

This affects the _collectorId attribute, which is different for every new collector registration.

To prevent this, remove the extensions.sumologic.clobber property or set it to false.

Cannot start reading file logs from specific point in time

The Filelog receiver currently supports only two modes of reading local files:

  • start_at: beginning: Ingest the whole file from the beginning, or
  • start_at: end: Only ingest newly added lines.

The start_at property is common to all the files read by the receiver - it cannot be set to end for some files and to beginning for other files. Note that this can be worked around by creating two separate Filelog receivers, one reading from the beginning and one reading from the end.

The other problem is that it is not currently possible for Filelog receiver to start reading files at a specific point, or only read files created or modified after a specific point in time.

There is currently no workaround for this.

Multiple multiline logs are sometimes concatenated

When using the Filelog receiver with a multiline configuration, multiple consecutive log lines can sometimes get concatenated into a single multiline log, even if the multiline configuration should cause these log lines to be split into mutliple separate logs (multiline or not). This issue is caused by a faulty implementation of flushing the multiline buffer in the OpenTelemetry log collection library.

A change that fixes this has already been prepared and should be available in an upcoming release.

There is no workaround for this issue, but setting the force_flush_period property to a higher value (e.g. 10s) should cause the issue to occur less frequently or not at all.