Monitor your Foxglove edge site.
The edge site performs periodic pings to the Foxglove API control plane to report health. Organization admins can see this status on the Sites settings page.
The edge site service exposes processing and memory statistics on a Prometheus metrics scraping endpoint on port 6002.
Check out the Prometheus documentation for more information.
recordings_indexed– Number of recordings indexed by the service
recordings_imported– Number of recordings imported to a Primary Site
recordings_deleted– Number of recordings deleted from the edge
recordings_gigabytes– Volume in recordings (GB)
imports_gigabytes– Volume of imported recordings (GB)
import_time_seconds– Histogram of import times to the Primary Site
import_throughput_mbps– Histogram of upload throughput to a Primary Site, when uploads are in progress
recording_summary_read_ms– Histogram of timings for reading MCAP summary sections from storage
recording_stat_ms– Histogram of timings for calling "Stat()" on MCAP files in storage
recording_index_ms– Histogram of timings for indexing recordings in the sqlite database
deletion_mark_ms– Histogram of timings for marking recordings for garbage collection in the index
deletion_sweep_ms– Histogram of timings for physical deletion of recordings from disk
control_plane_patch_recordings_deleted_ms– Histogram of timings for reporting garbage collections to the control plane
control_plane_post_recordings_ms– Histogram of timings for the (async) report of recording availability to console
recording_upload_ms– Histogram of timings for upload of a recording file to the configured inbox bucket
control_plane_patch_recordings_imported_ms– Histogram of timings for batched status updates issued to console after submission of recordings
pending_import_request_poll_ms– Histogram of timings for requests against Foxglove for pending import requests
database_recordings_count: Count of recordings indexed in the database.
database_topics_count: Count of unique topics/schemas indexed in the database.
database_recording_reports_count: Count of recording reports indexed in the database.
database_size_bytes: Total size of database in bytes.
To see a list of available metrics in your Terminal, use the
kubectl command to forward the Prometheus port:
$ kubectl port-forward -n foxglove service/edge-controller 6002:6002
$ curl -X GET http://localhost:6002/metrics
Slow query logging
Slow query logging in the edge site can be enabled by supplying the following environment variables. Using these should have negligible performance impact.
SLOW_QUERY_LOG_MIN_DURATION: A positive integer indicating the minimum execution time in milliseconds for a query to be considered "slow". Queries slower than this will be written to logs.
SLOW_QUERY_LOG_PARAMETERS: A boolean indicating whether parameters should be logged, defaulting to "false".
These environment variables can be supplied in the "edgeController.env" section of the helm chart.