We scan this tool’s changelog a few times a day and keep the changes that would alter your agent’s next action: 78 since January 2026, each linked to its source.
gcloudWhen creating or updating a transfer config for Dataplex Entry Group integration, use `--dataplex_entry_group`; to list configs with Dataplex configuration metadata, filter with `--filter="metadata_destination.dataplex_configuration:*"`.
bq mk --transfer_config --data_source=<DATA_SOURCE> --dataplex_entry_group=<ENTRY_GROUP>When metadata comes from standard input, pass `-` as the file path in `--metadata-from-file`. Use `KEY=-` on any of the four listed Compute commands.
printf '%s' '<metadata-value>' | gcloud compute instances add-metadata <INSTANCE_NAME> --zone=<ZONE> --metadata-from-file=<KEY>=-When creating a certificate with a generated key, set `--key-algorithm` to choose the client-side key algorithm. The documented values are `rsa-2048`, `rsa-3072`, `rsa-4096`, `ec-p256`, and `ec-p384`.
gcloud privateca certificates create <certificate-id> --issuer-pool=<pool> --issuer-location=<location> --generate-key --key-output-file=<key.pem> --key-algorithm=ec-p256 --cert-output-file=<cert.pem> --dns-san=<domain> --use-preset-profile=leaf_server_tlsWhen creating an HTTP or App Engine task with alpha or beta gcloud, pass the task-level retry flags to set that task’s retry behavior. For HTTP tasks, the documented flags are --max-attempts, --max-backoff, --max-doublings, --max-retry-duration, and --min-backoff; the App Engine reference does not document these flags.
gcloud alpha tasks create-http-task --queue=<QUEUE> --url=<HTTP_URL> --max-attempts=<N> --max-retry-duration=<DURATION> --min-backoff=<DURATION> --max-backoff=<DURATION> --max-doublings=<N>When you need to create multiple Cloud Tasks from a JSON or YAML file, run `gcloud alpha tasks batch-create` with the target queue and input file.
gcloud alpha tasks batch-create --queue=<queue> --tasks-from-file=<tasks.yaml>When deleting multiple Cloud Tasks, pass the queue and task input file to `gcloud alpha` or `gcloud beta tasks delete`; optionally write failed or skipped tasks to an output file. The input file can list task IDs/names one per line or use JSON/YAML format.
gcloud beta tasks delete --queue=<queue> --from-file=<tasks-file> --failed-tasks-file=<failed-tasks-file>To create bulk Compute Engine instances with selected machine types and a minimum CPU platform, use the beta bulk-create command with `--instance-selection` or `--instance-flexibility-policy`.
gcloud beta compute instances bulk create --name-pattern=<name-pattern> --count=<count> --region=<region> --instance-selection="name=<selection-name>,machine-type=<machine-type>,min-cpu-platform=<platform>"When creating a Dataproc cluster in a region with instances allowed across multiple zones, run `gcloud dataproc clusters create <cluster> --region=<region> --multizone`. For a workflow template’s managed cluster, run `gcloud dataproc workflow-templates set-managed-cluster <template> --region=<region> --multizone`.
gcloud dataproc clusters create <cluster> --region=<region> --multizoneTo configure HTTP/2 for a Cloud Run instance, deploy it with `--use-http2` to enable HTTP/2 or `--no-use-http2` to disable it.
gcloud alpha run instances deploy <INSTANCE> --image=<IMAGE> --use-http2For a dense vector search that needs a query-time recall target, pass `--dense-scann-target-recall=<value>` to the search command; the value must be between 0 and 1 inclusive.
gcloud vector-search collections data-objects search --project=<PROJECT_ID> --location=<LOCATION> --collection=<COLLECTION_ID> --vector-search-field=plot_embedding --vector-from-file=<SEARCH_VECTOR_FILE> --use-index=<INDEX_ID> --dense-scann-target-recall=0.95 --top-k=10