/storage/log/vcops/log/collector.log contains the following error meesagesCaused by: java.net.SocketTimeoutException: SocketTimeoutException invoking https://XXXX/api/vApp/vm-...: Read timed outWARN [Collector worker thread XX] com.integrien.alive.collector.CollectorWorkItem3.run - exceeded collect + process time: XXXXXX air={resId=XX, resUuid=XXX, resName=XXX, resKind=VCloudAdapter Instance, adKind=vCloud}, discarded=false, interval=300WARN [Adapter scheduler] com.integrien.alive.collector.DataCollectWorkItem.checkForRunCycle - Cycle = XXX MISSED for CollectorWorkItem XX, last cycle XXX air={resId=XX, resUuid=XXX resName=XXX, resKind=VCloudAdapter Instance, adKind=vCloud}, discarded=false, interval=300XX.XX.XX.XX - - [XX/XXX/XXX:XX:XX:00 +0000] "GET /api/vApp/vm-XXXXX/metadata HTTP/1.0" 200 XXXX "-" "vcd-client" XXXXAria Operations 8.18.x
Management Pack for vCloud Adapter 8.18.x
The adapter was configured to run resource-intensive sub-collectors (Advanced Metrics and Metadata) every 5 minutes, instead of the default interval of 1440 minutes. As a result, every collection cycle became significantly more resource-intensive.
In addition, the increased frequency of API requests generated substantial load on the vCD environment. As response times increased (e.g., 2–6 seconds per request), some collection cycles exceeded the 5-minute execution window.
Overall, overly aggressive collection interval settings led to excessive load, resulting in timeouts and dropped collection cycles.
Change the following Advanced Settings on vCD adapter instance:
ADVANCED_METRICS_COLLECTION_INTERVAL - 1440 (Default)
METADATA_COLLECTION_INTERVAL- 1440 (Default)
With these values, the metadata and advanced-metrics sub-collectors will only run once in every 288 cycles (at 1440 min) instead of in every cycle.