Skip to main content

Capture edge 2.1.10

Features​

Note: Maintenance releases typically avoid new features but may include small, low-risk enhancements.

  • Analytics: In a measurement, you can now choose by which trigger the detail data of this measurement is collected.
  • CloudManager: The version of the services is now displayed on the Status page of the CloudManager UI. Clicking the version shows more detailed information—the response of the /api/version call.
  • CloudManager: Added endpoint to get all logs as a zip: localhost:5004/api/log/downloadlogs. On Windows and Linux this zips all Capture log files on disk. On Docker it downloads the log files from the different services and merges it in one zip.
  • CloudManager: Added logs download link on the status page of the local CloudManager UI.
  • Collector: Keep data points in memory when connection to local db isn't ready yet or lost. By default we keep up to 50MB per collect type and database. This can be modified with the custom config property (on the local db) WriteQueueMaxSizeMB.
  • Docker: Included tools ps and procdump in the container images for easier debugging.
  • Syncer: Added endpoint to get the sync progress information: localhost:5001/api/syncstatus
  • UDP: Added support for the collect type Any.

Improvements​

  • General: Improved logging, performance, and automated test coverage.
  • General: Console logging is now disabled by default on Windows to reduce CPU usage. Can be enabled again by setting the CAPTURE_LOG_TO_CONSOLE environment variable to true.
  • General: Included commit hash in every log file for traceability.
  • ADS: Reconnects when no subscriptions can be made.
  • ADS: Improved CPU performance.
  • CloudManager: Giving identity.guid precedence over the database value.
  • Collector: Automatically removes incorrect http:// or https:// prefixes in Redis addresses.
  • Collector: Improved Influx connection resilience.
  • Collector: Improved CPU performance
  • Collector: The DataParseOptions custom option, previously only available for RedisPubSub, is now extended to all connections receiving strings.
  • Collector: Cadvisor now also supported on systems where containers are managed by systemd (in cgroup under /system.slice/docker-, instead of /docker/).
  • Collector: OnFollow child variables are collected first, so we have the last value before the parent variable triggers.
  • Influx: Drops NaN and Infinity double field values, as Influx does not support them.
  • Syncer: Prevents retrying file uploads to the cloud when they are rejected due to size limits.
  • Syncer: Delay next sync iteration if not enough data is returned by the local DB. Default: 500ms. Can be configured with the custom sync settings property BackoffWhenNotEnoughDataMs.
  • Syncer: Improved CPU performance
  • Syncer: Influx tag keys are cached in memory to reduce the number of Influx requests.

Fixes​

  • ADS: TwinCAT2 mappings starting with a . were not parsed correctly.
  • ADS: Properly handles multiple concurrent ADS routes on Linux.
  • ADS: Expands the array indexer on NonFlattedStructArray correctly.
  • ADS: Some reported issues were not cleared when the issue is resolved.
  • CAdvisor: Fixed incorrectly parsed numbers when they are too marge to fit in a signed long.
  • CloudManager: Prevents Capture services from restarting when a config is pushed while an update is in progress.
  • CloudManager: Fixed problem where not all update logs where sent to the cloud.
  • Collector: String values (e.g., from MQTT or Redis) containing numbers outside the bounds of an Int32 were not correctly collected.
  • GPS: Improved data collect frequency by avoiding unnecessary client disconnects.
  • Influx: Prevents unnecessary restarts on data formatting errors.
  • Influx: NaN and Infinity values are dropped before writing to Influx, as Influx doesn't support these values.
  • MQTT: Force stop MQTT thread when reconnecting, avoiding CPU leak
  • MQTT: Properly shuts down MQTT threads on reconnect to avoid CPU leaks.
  • Syncer: Fixed repeated transmission of points with the same timestamp until new points arrive. This occurred when points with a given timestamp were added to the local database after other points with the same timestamp had already been synced.
  • Syncer: Aborts the synchronization of a data packet if it contains no tags, indicating that Influx is not returning correct data.
  • Syncer: Improved thread safety in LiteDbClient.
  • Syncer: Avoid receiving double slave infos, so we don't lose a measurement in the sync queue.
  • UDP: Fixed an issue where points with different timestamps were incorrectly merged.

Known Issues​

  • Collector: Timer and Hybrid variables generate points at the lowest Timer/Hybrid interval of the connection, instead of the interval defined in the variable. Fixed in 2.1.11.
  • Collector: Influx reconnects resulted in CPU leaks. A high number of Influx reconnects could result in high CPU usage of the collector. Fixed in 2.1.11.

Noticed a problem with this version? Contact us at capture-support@vintecc.com.

Release created on: 07 April 2025 09:05:07